June 2016
Research Review No. 3110149017
Review of evidence on the principles of crop nutrient management
and nutrition for grass and forage crops
Paul Newell Price1, Kate Smith2 and John Williams2
1ADAS
Gleadthorpe, Meden Vale, Mansfield, Nottinghamshire, UK, NG20 9PD
2ADAS
Boxworth, Battlegate Road, Boxworth, Cambridgeshire, UK, CB23 4NN
This review was produced as part of the final report of an 8.5 month project (3110149017) which
started in September 2015. The work was funded by AHDB under a contract for £98,669. There
were six work packages. This report reviews findings from WP3 – Grass and forage crops.
While the Agriculture and Horticulture Development Board seeks to ensure that the information contained within this document is
accurate at the time of printing, no warranty is given in respect thereof and, to the maximum extent permitted by law, the Agriculture and
Horticulture Development Board accepts no liability for loss, damage or injury howsoever caused (including that caused by negligence)
or suffered directly or indirectly in relation to information and opinions contained in or omitted from this document.
Reference herein to trade names and proprietary products without stating that they are protected does not imply that they may be
regarded as unprotected and thus free for general use. No endorsement of named products is intended, nor is any criticism implied of
other alternative, but unnamed, products.
CONTENTS
1. ABSTRACT ......................................................................................................................... 3 2. INTRODUCTION ................................................................................................................. 4 2.1. 3. METHODOLOGY ................................................................................................................ 6 3.1. 3.1.1. 4. Aims and objectives .............................................................................................. 5 Online searches ..................................................................................................... 6 Screening the results ........................................................................................ 6 3.2. Projects and data ................................................................................................... 7 3.3. Deliverables ............................................................................................................ 7 RESPONSE OF GRASS TO APPLIED NUTRIENTS ......................................................... 9 4.1. Response of grass to nitrogen ............................................................................. 9 4.1.1. Response of different grass species or grass-legume mixes .......................... 15 4.1.2. Sward age ....................................................................................................... 20 4.1.3. Conclusions for grassland N recommendations .............................................. 21 4.2. 4.2.1. 4.3. 4.3.1. 4.4. 4.4.1. 4.5. Response of grass to sulphur ............................................................................ 22 Conclusions for grassland S recommendations .............................................. 24 Response of grass to phosphate and potash ................................................... 24 Conclusions for grassland phosphate and potash recommendations ............. 29 Response of grass to lime .................................................................................. 30 Conclusions for liming recommendations ....................................................... 30 Impact of grazing upon herbage production ..................................................... 30 4.5.1. Grazing height ................................................................................................. 33 4.5.2. Treading by livestock ...................................................................................... 34 4.5.3. Grazing returns ............................................................................................... 34 4.5.4. Conclusions for grazing and grassland management recommendations ....... 36 4.6. Grass and silage quality ...................................................................................... 37 4.6.1. Effect of nitrogen rate ...................................................................................... 37 4.6.2. Micronutrients.................................................................................................. 40 4.6.3. Effect of Biostimulants ..................................................................................... 41 4.6.4. Conclusions for grass and silage quality recommendations ........................... 42 5. GRASSLAND N RECOMMENDATION SYSTEMS .......................................................... 43 5.1. Fertiliser N recommendations in Scotland – SRUC Technical Note 652 ........ 43 5.2. Fertiliser N recommendations in the Republic of Ireland (ROI) ....................... 44 5.3. Fertiliser N recommendations in England and Wales – the “Fertiliser Manual
(RB209)” ........................................................................................................................... 46 6. 5.4. Comparison of N recommendation systems ..................................................... 49 5.5. Options for presenting nitrogen recommendations for grassland ................. 54 RESPONSE OF FORAGE CROPS TO APPLIED NUTRIENTS ...................................... 55 6.1. Response to nitrogen .......................................................................................... 55 6.2. Response to sulphur ........................................................................................... 59 6.3. Response to phosphate and potash .................................................................. 60 6.4. Conclusions for forage crop fertiliser recommendations ................................ 61 7. GAPS IN KNOWLEDGE AND FUTURE RESEARCH REQUIRED .................................. 62 8. CONCLUSIONS ................................................................................................................ 65 9. REFERENCES .................................................................................................................. 68 APPENDIX I – SEARCH TERMS AND NUMBER OF RESULTS.............................................. 78 APPENDIX II – OPTIONS FOR GRASSLAND N RECOMMENDATIONS ................................ 82 APPENDIX III - COMPANIES INVITED TO SUBMIT DATA AND/OR OPINIONS .................... 92 1.
Abstract
The Crop Nutrient Management Partnership was set up by AHDB to review and revise the
“Fertiliser Manual (RB209)” and produce a new “Nutrient Management Guide (RB209)” for release
in May 2017. The main aim of this work package was to review research carried out on nutrient
management in grass and forage crops to inform revisions of RB209. Literature searches were
carried out to identify research from the UK and northern continental Europe since 2009 to inform
revisions of the grassland and forage crop sections. Data was also gathered from a number of
relevant stakeholder groups and research projects.
Compared with trials carried out primarily in the 1980s and 90s, there is a relative paucity of data
on grass/clover and forage crop response to applied nutrients in England and Wales.
Nevertheless, recent evidence indicates that modern grass swards can produce more grass dry
matter (DM) yield than older swards at any given rate of applied nitrogen (N). This has important
implications for grass N recommendations, although further research is needed to confirm the
findings and to provide evidence for the response of grass to N when grazed. Declines in sulphur
deposition have increased the reliance of grass swards on sulphur mineralised from the soil and
applied as manufactured fertiliser or organic manure. There is increasing evidence to indicate that
sulphur should be applied to grass and forage crops in high risk situations (i.e. infrequent
applications of organic manure, on lighter soils and in higher rainfall areas).
Phosphate and potash recommendations are sound, but further work is needed on the response of
grass herbage to P fertiliser applications where soil P reserves are low or moderate; on identifying
and mapping high P-fixing soils in England and Wales; and on grass potash offtake values. The
importance of micronutrients for plant and animal health and options for supplying micronutrients to
crops and livestock also need to be covered in the new recommendations.
There was very little research or data made available on the response of forage crops to applied
nutrients, although there is some useful new data on forage maize. The evidence indicates that
only minor amendments are needed to forage crop recommendations.
Options have been provided for N recommendations that draw from the best aspects of the 7th and
8th editions. The existing sulphur, phosphate and potash recommendations from grazed and cut
grass have been reviewed. Recommendations for forage maize, Brassicas, whole-crop silage and
swede and kale grazed in situ have also been reviewed. Sections and appendices will be revised
and combined for a single entry on grass and forage crops for livestock farmers.
3
2.
Introduction
Defra’s Fertiliser Manual (RB209), 8th edition provided industry standard recommendations for
nutrient use for most agricultural crops including grassland. In the 8th edition, the structure of the
nitrogen (N) recommendations for grassland was completely changed from that in the 7th edition, at
the request of Defra to make the recommendations more applicable/relevant to the range of
livestock production systems. The new recommendations used a ‘systems’ approach with N
recommendations adjusted for the amount of grass production needed on a farm based on the
intensity of production and use of concentrates in the livestock system. Although this new
conceptual approach was widely regarded as a major improvement, the technical detail of the new
recommendations and the way they were presented in the 8th edition, was subject to much
discussion and there was no opportunity to test the new system on farmers and advisers before
publication.
The different livestock types, wide range of livestock and grassland management systems and
agro-climatic regions in England and Wales make the provision of grassland recommendations
challenging. The number and timing of cuttings and grazings varies within and between farms;
factors to consider which will vary from field to field and from region to region include:

Rainfall, altitude, temperature

Past management / soil type / Soil Nitrogen Supply (SNS)

Age of sward / clover content
It is important to account for nutrients recycled at grazing and limits to utilisation due to wastage
and spoilage of grass (Richards and Wolton, 1976; Richards et al. 1976 and Richards, 1978).
Improvements in the dry matter (DM) production potential of newer grass varieties (Chaves et al.,
2009; Sampoux et al., 2011; Wilkins and Lovatt, 2010) and the phosphate and potash offtake rates
of modern grass and clover varieties are another important consideration. New information on the
nutrient requirements of forage crops such as forage maize, Brassicas, whole-crop cereals and
root crops was needed for integration into livestock production systems. Nutrient management
guidance should also consider integration of feed advice in order to improve business profitability
and farm nutrient balances, which may be a vital component for sustainable production in the
future (AFRC, 1993; Scott, 2010).
Surveys carried out in 2012 and 2013 as part of Defra project IF01121 indicated that the 8th edition
grassland section was used by around 70% of grassland advisers, largely via hard copy, but only
13% of grassland farmers, with many finding it difficult to use. Field experiments carried out
between 2012 and 2014 indicated a greater DM yield response to N fertiliser from modern grass
4
varieties compared with trials carried out in the 1970s/80s that were used to underpin the 8th edition
recommendations. These findings identified a clear need for a technical update of the current
fertiliser recommendations for grassland and a thorough simplification of how the
recommendations are presented. Changes in forage crop practices with greater use of whole-crop
cereals and crops grazed in situ has also stimulated the need for an update of the fertiliser
recommendations.
This work package reviewed the current recommendations and additional relevant research with a
view to revising RB209 and producing the AHDB Nutrient Management Guide to be released in
2017. The new guide will provide practical, robust and clear information on nutrient management to
optimise economic production and minimise environmental impact by reduced nutrient loss.
2.1.
Aims and objectives
The main aim of the work package was to review research since 2009 on crop nutrition for the
main grassland and forage crops of England, Wales and Northern Ireland (N.I.) and based on the
findings, and where appropriate, to revise and amalgamate the grassland and forage crop sections
in the “Fertiliser Manual (RB209)” to produce new, clear, coherent, up to date, standalone and
scientifically robust recommendations. The main objectives were to:

Evaluate and review Defra and AHDB (and where applicable other UK) research undertaken.
since 2009 on the principles of crop nutrient management and nutrition for grassland and
forage crops.

Identify where changes to recommendations could be made.

Present changes in a format suitable for a future RB209 revision.
The specific detailed objectives were to:
i.
Use the review of evidence to develop new N recommendations for grassland.
ii. Provide options for presenting nitrogen recommendations.
iii. Review phosphate (P2O5) and potash (K2O) recommendations for grazed and cut grass and
update recommendations where necessary.
iv. Update forage crop recommendations, including forage maize, cereals grown for whole-crop
silage (with and without legumes), Brassicas, crops grazed in situ (swede and kale).
v. Amalgamate grassland and forage crop sections and related appendices into a single
complete section.
vi. Identify gaps in knowledge & future research required.
5
3.
Methodology
Relevant data and information from industry contacts (Appendix III) and from Defra and AHDB
projects was collated and reviewed for evidence produced since 2009. In addition, ‘Web of
Science’ searches were carried out using keyword combinations to find relevant UK-affiliated
research on crop nutrient management carried out since 2009.
3.1.
Online searches
‘Web of Science’ is a web-based scholarly research database and search facility that provides
access to bibliographic information such as the Science Citation Index
(http://apps.webofknowledge.com). Combinations of keywords, search strings and boolian
operators were used to reduce bias and provide more focused and productive results. The
geographical range was limited to United Kingdom and Western European affiliated research
papers, however this occasionally included research carried out elsewhere (e.g. New Zealand),
and if relevant this evidence was included in the review. Where searches returned a large volume
of papers, the geographical range was limited to the United Kingdom.
The database and search terms used, along with the number of hits (Appendix I) and a full list of
the evidence produced was recorded. Two screenings were then carried out to refine the search
results and identify relevant research papers.
3.1.1.
Screening the results
The first screening used the title of the evidence to check its relevance to addressing the aims and
objectives. The second filter used the full abstract to check for relevance and to select research
papers that would be used in the review. This resulted in a final list of 35 research papers.
The quality of the collated data was assessed for scientific rigour and relevance; field trials lacking
robust scientific protocols that did not include the use of replicated treatments in a randomised
block design or some other adequate replication methodology to take account of spatial variation in
soil properties (and other crop growth determining factors), standard error and variance were not
included in the review. All data sourced as part of the review were evaluated to assess the quality
of the data.
Where the evidence indicated a need to change fertiliser recommendations this is specified in the
following sections. Revisions will be made in an associated Appendix, which will be produced once
the selected option for grassland N recommendations (from those provided in this report) is clear.
The review clearly states where updates cannot be made due to a lack of robust or up to date
evidence.
6
3.2.
Projects and data
Evidence and data was collated from AHDB Dairy and AHDB Beef & Lamb research partnership
outputs and other grass and forage crop nutrient management projects, including:

AHDB Cereals & Oilseeds project 3699 – ‘Modern triticale crops for increased yields,
reduced inputs, increased profitability and reduced greenhouse gas emissions from UK
cereal production’.

AHDB Dairy project 74316 ‘Assessment of varietal characteristics important to low and zero
inorganic nitrogen input herbage production’.

Dale, A., Aubry, A., Ferris, C., Laidlaw, S., Bailey, J., Higgins, S. and Watson, C. (2013).
Critique of RB209 8th Edition Grassland nitrogen recommendations for dairy systems. AFBI
Report. 171pp.

Defra project IF01121 ‘Validation of Fertiliser Manual (RB209) recommendations for
grassland’

Defra project KT018 ‘Nutrient management decision support systems process
improvement’.

Defra project LS3650 ‘Utilise genetic variation within and between improved grass
populations to improve the sustainability of UK grassland’.

Wilkins, P.W. and Lovatt, J.A. (2010). Gains in dry matter yield and herbage quality from
breeding perennial ryegrass. In: Grasses for the Future. M. O’Donovan and D. Hennessy
(eds.) Proceedings of an International Conference, Cork, 14-15 October. Teagasc. 43-50.
3.3.
Deliverables
These included:

A comprehensive simplification of the recommendations (in terms of approach and
presentation), incorporating an appropriate level of precision while retaining scientific rigour
to take account of important factors such as livestock production intensity and system
potential.

Options for presenting N recommendations, including grass silage crop recommendations
and consideration of the influence of Grass Growth Class (GGC) and Soil Nitrogen Supply
(SNS).

Production of typical ranges of grass DM yield produced from different levels of N fertiliser
use.
Grass DM response to applied N
Assessment of grass N response incorporated findings from Northern Ireland (N.I.), Defra project
IF01121 ‘Validation of fertiliser manual (RB209) recommendations for grassland’ and from other
7
fertiliser N response plots used in projects such as DC-Agri (Defra/ WRAP, WRAP Cymru and Zero
Waste Scotland funded project OMK001-001) and Defra project AC0116 and the Innovate UK
‘Grass sense’ project.
New grassland N response data was compared statistically with IF01121 outputs (and the N
response curves generated for the 8th edition) and the overall response to differential rates of N
fertiliser assessed using ‘linear + exponential’ curve fitting and analysis of variance.
In addition to updating the N recommendations, we have reviewed the evidence to support
maintaining the current sulphur recommendations, including evidence from N.I. supporting the
possible need for additional sulphur applications ahead of first cut silage crops.
Options for presenting N recommendations
Outputs and experience from Defra project IF01121 surveys, case studies and focus groups have
been used to produce options for presenting N recommendations in a clear and simplified format
while retaining scientific rigour and an appropriate degree of precision. Options for presenting N
recommendations include:

Methods to account for different levels of livestock production intensity and system
potential.

Options for assessing the amount of fertiliser N to apply at grazing.

Options for treating silage as a crop, taking account of growth stage at cutting, quality
targets and sward clover content.
The following section reviews recent research on the response of grass to nitrogen, sulphur,
phosphate, potash and lime; the effect of grazing on DM yield; and the effect of nutrient
applications on grass and silage quality. Section 5 reviews grassland recommendation systems in
the UK and Republic of Ireland (ROI). Recent research on the response of forage crops to applied
nutrients is reviewed in section 6. Sections 7 and 8 outline gaps in knowledge, future research
requirements and the main conclusions of the review. Options for presenting new grassland N
recommendations are provided in Appendix II.
8
4.
Response of grass to applied nutrients
4.1.
Response of grass to nitrogen
The energy model that underpins the nitrogen recommendations in the “Fertiliser Manual (RB209)”
used a nitrogen response curve that was based on the GM20, GM23 and GM24 trials carried out in
the 1970s through to the early 1990s. There is some evidence to indicate that modern grass
varieties have a greater DM yield response to N fertiliser applications compared with the GM20-24
trials (e.g. Chaves et al., 2009; Sampoux et al., 2011 and Wilkins and Lovatt, 2010). Camlin (1997)
compared yields of cultivars bred in 1980 and 1995 and estimated that grass DM yield improved by
about 0.5% per annum as a result of greater growth potential of the newer varieties. More recent
estimates have ranged from 0.3% (Chaves et al., 2009; Sampoux et al., 2011) to over 1% per
annum (Wilkins and Lovatt, 2010). Field experiments carried out between 2012 and 2014 (Defra
project IF01121) also indicated a greater DM yield response to N fertiliser applications of newer
grass varieties compared to trials carried out in the 1970s/80s. The IF01121 experiments included
10 sites assessed over three years (2012-14) with sward ages mostly ranging between one and
ten years; one higher altitude (> 300 m above sea level) sward was around twenty years old. This
compares with the 48 trials in the GM20-24 datasets that were repeated in successive years.
Furthermore, all the IF01121 experiments were carried out at sites that had been in grass for a
number of years, while the GM20-24 trial sites had previously been in an arable rotation. This is
likely to have resulted in contrasting levels of Soil Nitrogen Supply (SNS) with the IF01121 sites
supplying greater levels of mineralised N. Nevertheless, it was worth comparing the older and
newer datasets to determine whether differences in the DM yield response were statistically
significant.
The N response curves from the two datasets were compared using regression analysis and
parallel curve analysis. Linear + exponential curves (Y= A + B*(R**X) + C*X; where Y is the DM
yield and X is the N fertiliser rate) were fitted to both sets of data for N fertiliser rates ranging from
0 to 450 kg N/ha. The parallel curve analysis involved 4 stages:
i.
A single curve was fitted to all the data.
ii. Parallel curves were fitted to the two sets of data with different intercepts for the 2 curves; A
was allowed to vary with parameters B, C and R kept constant for the two curves.
iii. A, B and C were allowed to vary with R kept constant for both curves.
iv. All parameters were allowed to vary, i.e. the curves fitted were the best for each individual
dataset.
At each stage the sums of squares (s.s.) explained by the fit was calculated and the improvement
in fit determined (Table 1). The percentage variance accounted for was 64% at stage 1 and 84% at
9
Stage 2, providing a significant improvement in fit using the parallel curves compared with the
single curve. Stages 3 and 4 did not provide a significant improvement on stage 2, indicating that
the data can be fitted best by two parallel lines (Table 1 and Figure 1) with the DM yield response
from the IF01121 data around 3.9 t DM/ha above that for the GM20-24 data.
Table 1. Accumulated analysis of variance from parallel curve fitting to compare IF01121 and GM2024 N response data.
Change
Degrees of
freedom
Sum of
squares
Mean
square
Variance
ratio
P value
+ N (Stage 1)
3
1378.940
459.647
168.39
<0.001
+ Group (Stage 2)
1
433.208
433.208
158.71
<0.001
+ N.Group (Stage 3)
2
0.173
0.086
0.03
0.969
+ Separate non-linear (Stage 4)
1
0.025
0.025
0.01
0.924
Residual
120
327.557
2.73
Total
127
2139.903
16.85
20
18
Dry matter Yield (t ha‐1)
16
14
12
10
IF01121
8
6
4
RB209
2
0
‐100
0
100
200
300
400
500
Nitrogen application rate (kg N ha‐1)
Figure 1. Parallel curves fitted to the IF01121 and GM20-24 DM yield response data.
Wilkins and Lovatt (2011) found that modern perennial ryegrass varieties yielded 12-38% more
than older varieties (cv. Talbot and cv. S23) at the N fertiliser application rate tested of 385 kg
10
N/ha. By comparison, the IF01121 swards on average yielded 36% more than the older GM20-24
varieties at the same level of N use (Figure 1). This is at the upper end of the range measured by
Wilkins and Lovatt (2011). However, if modern grass varieties are more efficient in their ability to
convert N into dry matter one would expect the DM yield differential to be greater at higher N
fertiliser rates than at lower rates, i.e. the two N response curves should be divergent, with the
modern variety response curve having a steeper slope compared with the older variety response
curve. The fact that the difference in yield between old (GM20-24) and modern (IF01121) varieties
was similar at 0 kg N/ha and 400 kg N/ha indicates that the difference between the two datasets
was more to do with SNS (or the comparative ability of the swards to exploit the nitrogen supplied
from the soil) than the relative fertiliser N use efficiency of modern and older perennial ryegrass
varieties. More work is needed comparing the N response curves of older and modern ryegrass
varieties in replicated field experiments at a number of sites using differential N rates. Therefore,
there is insufficient evidence from replicated UK grassland field experiments to justify a change to
the N response curves used to underpin the grassland N recommendations in the “Fertiliser
Manual (RB209)”.
N response by defoliation
When providing N recommendations for grassland it is important to consider the amount of N to
apply for each cut (of silage or hay) and grazing. The effects of grazing on herbage production is
covered in section 4.5. To investigate DM yield response by cut, data was collated from a number
of replicated, randomised block, field experiments using differential fertiliser N rates carried out in
England, Wales and N.I.:

Defra IF01121 – 4 cuts taken at 10 separate sites; first cut in late May/early June with
subsequent cuts at approximately six week intervals

AFBI Hillsborough Estate, County Down – 3 cuts taken at 4 separate sites; first cut in midMay/early June with subsequent cuts at six to eight week intervals

Defra AC0116 – 3 cuts taken at two separate sites; first cut in mid-May/early June with
subsequent cuts at six to eight week intervals

Innovate UK, Grass Sense project – 2 cuts taken at three separate sites; first cut in
May/June and second cut in July/August
Linear + exponential curves were fitted to the data and the optimum N rate calculated using a
break-even ratio of 10:1 (Chadwick and Scholefield, 2010). Meta-data on GGC, SNS and sward
age was also collated to investigate the effect of these factors on DM yield response. The
experimental protocols for first cut silage were the most consistent in terms of the defoliation timing
and yet there was a remarkable degree of variation in DM yield response (Figures 2 to 4). Nopt10
(The fertiliser N rate at economic optimum yield, based on a break-even ratio of 10:1) ranged from
11
62 to >170 kg N/ha, while the DM yield ranged from 2.3 to 7.7 t DM/ha (Table 2). The variation in
DM yield at each N fertiliser rate was mainly a function of SNS (Figures 2 to 4).
Table 2. Ranges of nitrogen (N) optimum fertiliser rate at a break even ratio of 10:1 by cut and for the
whole season and number of observations for which N-optimum was reached. Note: for the
remaining observations N-opt was either not reached or exceeded the highest N-rate.
1st Cut
2nd Cut
3rd Cut
4th Cut
N-opt (kg N/ha)
62 to 170
47 to 147
30 to 91
15 to 38
Whole
season
110 to 406
Dry matter yield at N-opt (t/ha)
2.3 to 7.7
2.4 to 5.9
1.2 to 5.7
0.3 to 2.9
7.0 to 16.9
Number of observations N-opt
16
19
8
3
19
22
22
16
10
22
reached
Total number of observations
Dry matter yield (t/ha)
First cut
Second cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
0
250
50
100
150
200
250
200
250
Fourth cut
Third cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
0
250
50
100
150
Nitrogen fertiliser rate (kg/ha)
Figure 2. Fitted (linear + exponential) yield response curves by cut; colour coded by Soil Nitrogen
Supply (SNS) Index: Low, Moderate and High. Diamonds represent grass yields at nitrogen optimum
at a break even ratio of 10:1.
12
Dry matter yield (t/ha)
First cut
Second cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
250
0
50
Third cut
100
150
200
250
200
250
Fourth cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
0
250
50
100
150
Nitrogen fertiliser rate (kg N/ha)
Figure 3. Fitted (linear + exponential) yield response curves by cut; colour coded by Grass Growth
Class: Poor/Very poor, Average, Good and Very good. Diamonds represent grass yields at nitrogen
optimum at a break even ratio of 10:1.
The Nopt10 for second cut grass varied from 47 to 147 kg N/ha, while the DM yield ranged from 2.4
to 5.9 t DM/ha. For third cut grass, Nopt10 ranged from 30 to 91 kg N/ha and the DM yield at 75 kg
N/ha ranged from 1.2 to 5.7 t DM/ha. For fourth cut, Nopt10 ranged from 15 to 38 kg N/ha and the
DM yield ranged from 0.3 to 2.9 t DM/ha. Given the large degree of variability in DM yield response
to applied N and in Nopt10, and to optimise yield and N fertiliser use efficiency, the N
recommendations at each cut should take into account both the potential productivity of grass
swards at higher rates of applied N and the N level at which some swards approach a plateau in
terms of their DM yield response. In any particular field and year, it is impossible to predict the
shape of the N response curve without applying differential N rates to plots set up for this purpose,
so N recommendations should reflect the risk of applying N beyond maximum yield (based on
recent data) and therefore the risk of low N fertiliser use efficiency at higher N rates.
13
Dry matter yield (t/ha)
First cut
Second cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
250
0
50
100
150
200
250
200
250
Fourth cut
Third cut
9
9
8
8
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
0
50
100
150
200
0
250
50
100
150
Nitrogen fertiliser rate (kg N/ha)
Figure 4. Fitted (linear + exponential) yield response curves by cut; colour coded by sward age: less
than 5 years old and greater than 5 years old. Diamonds represent grass yields at nitrogen optimum
at a break even ratio of 10:1.
Higgins et al. (2012) in a field study in N.I., assessed the effect of annual applications of N fertiliser
applications (at 0, 75, 150, 225 and 300 kg N/ha) as calcium ammonium nitrate on grass (the
sward was predominantly meadow grass and ryegrass) productivity and soil chemical properties.
Overall, the study found that there was a highly significant (P <0.001) effect of N rate on grass dry
matter yields (Table 3), with mean annual (all 3-years) yields reaching 11.5 t/ha at a N application
rate of 300 kg N/ha/yr.
14
Table 3. Mean annual total (three cuts) grass dry matter (DM) yield (t/ha) in response to nitrogen (N)
fertiliser rate (highly significant (P <0.001) effect), for pelleted and ground lime combined (no
significant difference between different liming materials). Taken from Higgins et al. (2012).
Mean annual DM yield (t/ha)
/
4.1.1.
N-rate (kg N/ha yr)
2007
2008
2009
0
9.01
4.94
4.71
75
10.04
6.87
7.35
150
11.29
8.91
9.91
225
12.09
10.61
11.26
300
11.72
11.15
11.76
LSD
0.62
0.54
0.48
Response of different grass species or grass-legume mixes
A number of studies have investigated the response of different grass species or mixes to N
fertiliser inputs. For example, Dale et al. (2015), in a field experiment located in N.I., assessed the
response of seven forage types, to cattle slurry applied at 4 different rates (supplying 0, c.100,
c.200 and c.300 kg total N/ha/yr) in 3 split applications by trailing shoe. Average DM yield response
to slurry was 15.6 kg DM kg-1 N. The results suggest that slurry applied to swards containing
legumes on soils with a high P-content will have a lower DM response to slurry N, resulting in a
lower slurry-N recovery than on swards of perennial ryegrass or cocksfoot-dominated low-input
mixtures. To maximise slurry N-recovery, it was recommended that applications after first-cut
where growth rate is likely to be slow or where there is a high legume content should be avoided.
AFBI carried out a series of grass mixture trials from 2003 to 2005 in Loughgall, Co Armagh, N.I.
Ten mixtures were included in the trial comprising various proportions of tall fescue, timothy,
perennial ryegrass, cocksfoot and meadow fescue. The trials were maintained for three years
using a replicated (three times) design of the following treatments:
i.
Organic management: no manufactured N fertiliser applied, with white clover included in
the mixture and an application of phosphate and potash to simulate a slurry application
ii. Low Input: 170 kg N/ha, 99 kg P2O5/ha and 145 kg K2O/ha per annum
iii. High Input: 360 kg N/ha, 294 kg P2O5/ha and 240 kg K2O/ha per annum
In 2004, the mean DM yield under organic management was 36% of the High Input management.
White clover progressively dominated the swards and only 3 cuts were achieved. The highest DM
yield under organic management of 6.6 t/ha was from a grass mixture comprising 50% Bareno
Brome, 30% cocksfoot and 20% late tetraploid perennial ryegrass.
15
The highest yield under the Low Input treatment was 14.1 t DM/ha from a mixture comprising 20%
late tetraploid perennial ryegrass, 20% timothy, 20% meadow fescue, 20% tall fescue and 20%
cocksfoot. The highest yield under the High Input treatment was 17.2 t DM/ha from a mixture
comprising 50% brome, 20% late perennial ryegrass and 30% tall fescue. The most persistent and
productive mixtures over a 3 year period were ‘Barmix’ type mixtures, with a component of deeprooted species, tall fescue and cocksfoot. It was also found that cocksfoot tended to dominate
when its component in the sward was increased. The highest yields compare favourably with yields
achieved using perennial ryegrass dominated swards (Figure 1).
The main advantages of grass-legume swards is that they reduce reliance on N fertiliser inputs,
tend to produce higher yields compared to growing component species alone at low N fertiliser
rates, produce more balanced feeding values and increase nutrient-use efficiency. Disadvantages
of forage legumes include lower persistence than grass under grazing, unpredictable nature of Nfixation, lower productivity than systems reliant on manufactured N fertiliser (Humphreys et al.,
2009), risk of livestock bloat, difficulties with preservation when ensiling and with maintaining
optimum legume proportions (Phelan et al. 2014).
Høgh-Jensen and Schjoerring (2010) in a field experiment located in Denmark, assessed the
impact of NPK fertiliser applications (see Table 6for rates) made during the first week of April on
yield and N-offtake of ryegrass and white clover grown separately or as a mixture, when the
swards were 1 and 2 years-old. It was found that without the addition of manufactured N fertiliser,
ryegrass-clover mixtures (sown to achieve a target composition of 1/3 clover to 2/3 ryegrass)
consistently out yielded pure stands of either clover or ryegrass in terms of DM production.
Furthermore, ryegrass grown with clover, accumulated substantially higher amounts of N, P, and K
than ryegrass in a pure stand. The growth of white clover was significantly depressed by N
application, particularly when P and K were also applied. In contrast, ryegrass yield responded to
increasing availability of N, P and/or K. The study demonstrated there is a complex interaction of
growth and nutrient acquisition when ryegrass and clover are grown as a mixture.
Suter et al. (2015), in a 3-year field study carried out across the EU, compared total N-uptake of
grass monoculture and grass mixtures containing varying amounts of legumes. At each site,
manufactured N fertiliser was applied at a standard rate across all plots, but differed (0 to 150 kg
N/ha) between sites reflecting differences in background productivity. Overall, it was found that in
all years N-uptake was significantly (P <0.05) greater in grass-legume mixtures (195 to 286 kg
N/ha/yr) than in grass monocultures (119 to 178 kg N/ha/yr). Nitrogen-uptake increased with
increasing proportion of legumes up to one third of sward composition. Grass mixtures containing
one third legumes attained similar to 95% of the maximum N-total acquired by any stand and had
16
57% higher N-total than grass monocultures. Furthermore, relative N gain by mixtures was not
related to site productivity, indicating that grass-legume mixes can provide increases in N-uptake
over grass monocultures in both low and high productivity sites, although legume covers and the
increased uptake associated with grass-legume swards was limited by temperature. This adaption
to site productivity can be explained by N2 fixation being largely regulated by the N-sink strength of
the whole system. The study also reported that N-losses from grass-legume swards can be lower
compared to fertiliser grassland systems due to three mechanisms: i) nitrogen fixed symbiotically is
stored within legume nodules and is therefore not freely available, ii) symbiotic N2 fixation is
downregulated if demand is low and iii) in grass-legume systems, grass species take up N-fixed by
legumes and from mineralised organic matter. The authors concluded that the use of grass-legume
mixtures can contribute to improving resource use efficiency in grassland systems over a wide
range of productivity levels, implying important savings in manufactured N fertiliser and significant
potential for climate change mitigation.
Nyfeler et al. (2011) reported that the acquisition of fixed-N was maximised in grass-legume mixed
swards compared to pure legume stands. The highest yields were achieved in mixed grass-legume
swards receiving low to moderate N-fertiliser inputs (50 to 150 kg N/ha) containing c.40-60 %
legumes. It was concluded that, increasing the proportion of grasses, increased both the proportion
of N-fixed contained in the legume species and the transfer of fixed-N to grasses; legumes are able
to regulate N2-fixation according to N-demand (or sink). Therefore, in order to maximise N-fixation
and N-acquisition by grass-legume swards, it is important to achieve the correct balance between
N-fertiliser inputs and the proportion of grass and legumes species in a sward.
Other authors have investigated the productivity and profitability potential of grass and grass-clover
swards. Within dairy systems, stocking rate and milk output tend to be higher on grass-based
systems receiving manufactured N fertiliser compared to grass-clover based systems. However,
grass-clover can support an annual stocking density of 2.15/ha and a milk output of 14 t/ha
(Humphreys et al., 2009). In a later study, Humphreys et al. (2012) assessed the profitability of
dairy production systems based on N fertilised grass and grass-white clover grassland, by
comparing data collected from system-scale studies carried out in Ireland between 2001 and 2009.
Ten fertilised grass systems stocked between 2.0 and 2.5 livestock units (LU) ha-1 with N rates
between c.170 and c.350 kg N/ha were compared with 8 grass-white clover systems stocked
between 1.75 and 2.2 LU/ha with N fertiliser inputs between c.80 and c.100 kg N/ha. The study
compared the profitability of systems with changes in manufactured N fertiliser and milk prices. It
was found that, in scenarios of high manufactured N fertiliser prices combined with intermediate or
low milk prices grass-white clover systems were more profitable than fertilised grass systems.
Given manufactured N fertiliser and milk prices at the time of the study (1990 – 2005), fertilised
grass systems were more profitable than grass-clover systems. However, with increasing
17
manufactured fertiliser costs relative to milk prices differences between the two systems became
less pronounced. The study concluded that, in the future, it is likely that grass-clover systems will
become a profitable alternative to grass systems relying on manufactured N fertiliser.
Gierus et al. (2012) in a field experiment located in Germany, assessed the DM yield production of
6 different forage legume species grown with perennial ryegrass compared to monoculture of
ryegrass (with or without slurry applied at 200 kg N/ha; supplying approximately 70 kg crop
available N/ha) in a 1 year-old sward. Overall, it was found that all legume-ryegrass mixes yielded
more than ryegrass (both with and without slurry application) grown as a monoculture. Among the
legume-grass mixes the lowest yields were achieved with ryegrass grown either with Birdsfoot
trefoil (mean c.6.3 DM/ha) or Caucasian clover (mean c.5.8 kg DM/ha), due to deficient
establishment of the legumes. The highest ryegrass yields were achieved when grown as part of a
white clover-ryegrass mix (c.4.4 kg DM/ha). However, maintaining the optimum legume content (of
40-60% of herbage dry matter) to achieve these benefits remains a major challenge. Consistent
with Humphreys et al. (2012), the findings support that grass-legume mixes yield well and can be
used to reduce costs and improve profitability when fertiliser prices are high and milk commodity
prices low.
King et al. (2012) compared the effect of N-fertiliser (0 and 125 kg N/ha) and harvest date on the
yield and chemical composition of five common grasses and red clover at Grange, Co. Meath,
Ireland. Data were analysed using regression analysis, allowing comparisons between species at
common growth stages whilst removing the confounding effect of differences in maturity between
species harvested on the same date. The main findings from the study were that Timothy grass
was the most productive in terms of dry matter yield, but the poorest in terms of digestibility, with
potential impacts on both animal and bioenergy production potential. Nevertheless, timothy has the
potential to provide cheaper feed per unit of DM than other grass species. However, Italian
ryegrass was the most suitable for ensiling due to its higher water soluble carbohydrate
concentration. While red clover had a higher mean DM digestibility and crude protein
concentration, but lower water soluble carbohydrate (WSC) concentration and higher buffering
capacity, which could compromise preservation during ensiling.
Høgh-Jensen and Schjoerring (1994), reported that N accumulation more than doubled in ryegrass
clover mixtures compared to ryegrass grown alone (Table 4). The study also reported increases in
dry matter and N accumulation when P fertiliser was applied, however K had no effect. The
symbiotic N2-fixation was determined in this experiment using 15N isotope dilution as described in
detail by Høgh-Jensen and Schjoerring (1994), using grass in pure stand as the reference (Fried
and Middelboe, 1977). The quantity of fixed N2 (BNF) in the harvested clover when grown alone,
18
was increased when P or K was applied. When clover was grown with ryegrass, harvested BNF
increased following P or K application, but only when no N fertiliser was applied (Table 4).
Table 4. Nitrogen (N) accumulation and amount of N fixed by clover in brackets in two production
years. Taken from Høgh-Jensen and Schjoerring (2010).
Year
Treatment N, P, K
(kg/ha)
1
0, 0, 0
232b
(201)ab
178ab
0, 0, 120
251ab
(201)ab
0, 20, 0
280a
0, 20, 120
2
Pure Clover
(kg/ha)
Clover mix (kg/ha)
Grass mix
(kg/ha)
Pure
Grass
(kg/ha)
(161)ab
49c
18c
148bc
(141)bc
36c
24c
(225)a
189a
(179)a
46c
20c
271ab
(223)a
191a
(181)a
58c
23c
120, 0, 0
238ab
(174)c
139c
(120)cd
110b
81b
120, 0, 120
259ab
(177)bc
94d
(86)e
126ab
96ab
120, 20, 0
277ab
(178)c
112cd
(103)de
136ab
90ab
120, 20, 120
271ab
(191)bc
83d
(77)e
142a
102a
0, 0, 0
163c
(142)cd
65d
(60)d
52e
11b
0, 0, 120
200b
(170)bc
131b
(124)b
62de
14b
0, 20, 0
213b
(179)ab
139b
(131)b
79c
16b
0, 20, 120
233a
(203)a
167a
(158)a
73cd
13b
120, 0, 0
169c
(115)d
69d
(56)d
119ab
74a
120, 0, 120
211b
(167)bc
108c
(95)c
114b
80a
120, 20, 0
217ab
(172)b
102c
(88)c
124ab
82a
120, 20, 120
232a
(190)ab
98c
(89)c
135a
82a
In a review of the potential of legumes to increase sustainability of ruminant production systems,
Peyraud et al. (2009) concluded that the yield value of grass-clover mixes (consisting of 30-80%
clover) can be equivalent to fertiliser N-inputs of 150 to 350 kg N/ha and productive grass-clover
mixes can fix 150 to 350 kg N/ha (Peyraud et al., 2009). This agrees with the potential nitrogen
supply values given in the “Fertiliser Manual (RB209)”.
The N response experiments set up as part of Defra project IF01121 included two sites in
Ceredigion (west Wales) and Devon (south west England) with 7% clover in the seed mix (w/w).
SNS at these sites would have been low without clover addition; making it possible to compare the
productivity of these sites with other low SNS sites without clover in Ceredigion and Shropshire
(Figure 5). Standard rates of manufactured N fertiliser were applied; 0 to 450 kg N/ha over four
cuts. Clover contents at both grass/clover sites were high by mid-season (>40%) and were
effective in increasing yield at lower manufactured N fertiliser rates, typically increasing grass DM
yield by 2-3 t/ha compared to swards without clover. The data indicates that these modern clover
varieties can fix around 100-150 kg N/ha over the growing season (which is at the lower end of the
19
range suggested by Peyraud et al., 2009) and that modern clover varieties can tolerate up to 240
kg N/ha in the form of manufactured N fertiliser over three cuts. Overall, the results demonstrate
the importance of clover in low to moderate output systems.
Figure 5. Grass dry matter (DM) yield response to manufactured nitrogen (N) fertiliser in two
grass/clover swards and two grass only swards in 2013. The curves were fitted to the data using a
linear + exponential equation. Data taken from Defra project IF01121.
The results, while interesting, are from a limited number of sites and do not negate the need for
further N response experiments on a selection of grass and grass-clover swards at the same site
or sites. Old and modern perennial ryegrass varieties could be included to assess contrasting DM
yield responses at differential N rates.
4.1.2.
Sward age
Sward age is often cited as an important factor determining grassland productivity (e.g. Frame and
Laidlaw, 2014; Cashman et al., 2016). However, there is limited evidence assessing the
productivity and quality of permanent grassland swards (i.e. swards greater than five years old),
and in particular how botanical composition of older swards affects grass yield and nutrient value,
and how this varies with environmental conditions (Michaud et al., 2015). In a two year study
(2009-2010) Michaud et al. (2015) assessed how production and forage quality varied with species
composition in 190 permanent grasslands located in France, spanning a range of soil, climatic and
management conditions. Michaud et al. (2015) found that over a wide range of environmental and
different management practices, vegetation characteristics (i.e. functional types e.g. legumes)
explained approximately half of the variance of forage quality and 20-40% of the variance of
20
biomass production. This view is supported by Baumont et al. (2012) who explained that the forage
value of permanent grasslands can in some cases be equal to that of sown grasslands.
In a study of ten perennial cultivars conducted over three years at Teagasc Moorepark, Fermoy,
Co Cork, Ireland, Cashman et al. (2016) found that DM yields under conservation management
(six mechanical defoliations from late March to mid-October) were c.18% higher on average in the
first year after sowing compared with the second year; and c.43% higher on average compared
with the third year. They concluded that the results from the first harvest year may not represent
the long-term performance of a grass sward; and also found that DM yield and sward density
decline faster when swards are under conservation management as opposed to grazing
management (Cashman et al., 2016).
Comparison of N response data from Defra project IF01121 and from the GM20, GM20, GM24
(cutting data only) and GF01 (permanent pasture treatments only) trials carried out from 1970 to
1993 indicate that modern grass varieties produce higher DM yields than older varieties at any
given level of N supply (section 4.1). Indeed, the data indicates that one to ten year old grass
swards with minimal invasion by less productive species can yield 15% to 36% more DM than
older varieties used in the trials that underpin the 8th edition grassland recommendations. This is
supported by the experiments carried out by Wilkins and Lovatt (2011). However, there was no
relationship between sward age and DM yield response within the IF01121 dataset (12 sites all
with one to ten year old swards), other than for a first year ley that out-yielded all other swards (cf.
Cashman et al., 2016), indicating that lower productivity relates to swards older than ~10 years,
swards with a significant production of less productive species, such as broadleaved weeds (e.g.
Rumex spp.) and meadow grass, and swards at higher altitudes (the altitude ‘cut off’ for
downgrading the Grass Growth Class by one in RB209 is 300 m). It is the combination of sward
age (older/newer varieties) and sward composition (less/more productive species and the influence
of clover) that determine the DM yield response. Grassland N recommendations should therefore
provide an indication of the typical range of DM yields that can be achieved at any given level of N
supply for well-managed younger and older swards.
4.1.3.
Conclusions for grassland N recommendations
Grassland N recommendations should aim to take account of the contrasting DM yield response of
older and newer perennial ryegrass varieties to applied nitrogen and the wide range of sward
composition in terms of the proportion of modern perennial ryegrass varieties that they contain.
However, there is limited evidence of how modern perennial ryegrass varieties respond to
differential rates of applied N and limited data on the age of swards or the proportion of modern
grass varieties in swards. For example, the IF01121 N response data was derived from a limited
number of sites and agro-climatic conditions, and it is difficult to assess how typical the swards
21
were of current grassland in England and Wales. Nevertheless, the IF01121 data provide some
indication of the growth potential of modern grassland swards, and could be used to guide
recommendations based on indicative DM yield ranges; to reflect the N response of older and
younger grass varieties and the varying composition of grass swards.
The review by Peyraud et al. (2009) of the contribution of legumes to N supply are in agreement
with the potential nitrogen supply figures for clover in the “Fertiliser Manual” (RB209). The values in
RB209 should therefore be retained. Nevertheless, there is a need for field experiments to
determine the N supply of modern clover varieties, comparing grass-clover swards with older and
more modern perennial ryegrass varieties at the same site or sites.
4.2.
Response of grass to sulphur
Sulphur (S) is an essential nutrient for grass growth and nutritional quality (Bailey, 2016). Grass
and particularly legumes have one of the highest requirements for S. Indeed, as sulphur deposition
reduced in the UK in the 1970s and 80s grass and oilseed rape were the first crops to show S
deficiency symptoms with associated impacts on yield.
Declining S deposition during the last 30-years and the decreasing use of S containing fertilisers
has led to an increase in S-deficient crops including grass and legumes. Indeed, S deposition no
longer supplies a significant proportion of crop S requirements (Webb et al., 2015). Tallec et al.
(2008) reported that the greater S requirement of legumes compared with grasses leads to interspecific competition between Trifolium repens (white clover) and Lolium perenne (perennial
ryegrass) under cutting, resulting in the replacement of clover by grass. Reduced N fixation by
clover therefore leads to decreased sward productivity in low input systems and the need for
greater inputs of manufactured N fertiliser. Reductions in S deposition could therefore result in a
reduction in the abundance of leguminous species in grassland swards.
Defra project SCF0308 (Webb et al., 2015) concluded that within grassland systems, the need for
S fertiliser appears to be greatest for grass swards cut more than once; and that the amount of S
currently being applied is insufficient to provide optimum yield. Brown et al. (2000) reported yield
increases of 35% on sandy soils and 11% on clay soils for swards cut 3 times per year and
fertilised with 400 kg N/ha. Largest yield responses have previously been detected from 2nd and
later cuts (Scott et al., 1983; Stevens and Watson, 1986; Brown et al., 2000).
The “Fertiliser Manual (RB209)” advises the application of 40 kg SO3/ha as a sulphate containing
fertiliser applied at the start of growth before each cut. However, only 10% of grassland soils
receive S fertilisers (Anon., 2014); and the average SO3 application to grassland (33 kg SO3/ha) is
less than the average application applied to arable crops (58 kg SO3/ha). Livestock manures are
22
applied to around 50% of grassland soils (Anon., 2014), providing useful quantities of S, but this
leaves half of the grassland area relying on S deposition and mineralisation of organic matter to
supply S to the growing crop. On lighter soils and in higher rainfall areas (prone to S leaching) this
could lead to S deficiency in grass and grass-clover swards.
Given the difficulty in predicting where crops may respond to S fertiliser and the declining role of S
deposition in meeting crop demand, Webb et al. (2015) conclude that the guidance proposed by
Cussans et al. (2007) should be incorporated into RB209, i.e. that S is applied, either as mineral
fertiliser or livestock manures, to all crops and grass grown on:

Sandy soils

Loamy and coarse silty soils in areas with > 175 mm overwinter rainfall

Clay, fine silty or peat soils in areas with >375 mm overwinter rainfall
Data from N.I. indicates that S deficiency can be prevalent at first cut even on heavy textured soils,
but is often corrected by cuts 2 and 3 (Bailey, 2016). Herbage analysis and interpretation using the
Diagnosis and Recommendation Integrated System (DRIS) provides an accurate indication of the
S sufficiency status of grass and hence the degree to which S supply is influencing grass DM yield.
The addition of organic manure had a small effect on grass S sufficiency status, but cannot be
relied on to supply the S needs of silage crops. Further work is therefore needed using S fertiliser
to produce a response curve. This has been identified as a research gap for the application of S
fertiliser before first cut (Bailey, 2016; Eriksen, 2009). Bailey (2016) recommended that S should
be applied routinely to all grass silage swards in early spring, at a rate of 35-40 kg SO3/ha, “since
this rate of application should eliminate all risk of S deficiency, with little risk of S over-supply even
on moderate texture soils, and in many cases obviate the need for further S applications later in
the season”. Some nutrient management systems recommend using the N:S ratio of grass herbage samples to
determine S deficiency. For example, SRUC Technical Note 652 indicates that a response to S
fertiliser is highly likely when the total N:total S ratio in herbage is greater than 16:1 (Sinclair et al.,
2013). This is supported by Mathot et al. (2009) who discuss the development of indicators for S
deficiency in grass using data collected from field and pot experiments carried out from 1986 to
2008, and propose a diagnostic tool based upon linear relationships linking the S and N content of
grasses. The dataset excludes swards containing more than 20% of legumes. Using the tool, grass
S nutrient status can be placed into one of four categories: certainly sufficient, probably sufficient,
probably deficient and certainly deficient. The authors recommend validating the tool with a large
independent dataset.
23
Sirius minerals provided data from an experiment carried out on a permanent grass sward on
sandy loam soil in Central England, assessing the yield response of second cut grass silage to
sulphur fertiliser application. Sulphur fertiliser applications increased grass dry matter yields by
c.0.8 t/ha compared to the untreated control (P< 0.01), with grass responding to application rates
of up to 20 kg/ha SO3.
4.2.1.
Conclusions for grassland S recommendations
The current recommendations appear sound, but the advice to apply S fertiliser in the situations
suggested by Cussans et al. (2007) could be included. Given the findings of Bailey (2016), it may
also be sensible to remove the sentence “Deficiency at first cut is less common but can occur on
light sand and shallow soils”, while being wary that over supply of S can be detrimental to animal
health.
4.3.
Response of grass to phosphate and potash
Potash supply should be proportional to the amount of nitrogen applied to grassland although
applications in spring prior to first cut silage (or hay) should be limited to 80-90 kg K2O/ha to
reduce the risk of ‘staggers’ (hypomagnesaemia). The aim should be to apply ‘maintenance
applications of potash fertiliser to balance the offtake in cut or grazed grass over the year (Sinclair
et al., 2013).
The yield response of grass to phosphorus (P) is more variable and soil analysis does not always
predict such variation (Valkama et al., 2016). For phosphate, grass DM yield responses are
generally significant when soil P reserves are low, although this is not always the case (Mahli et al.,
2009); while at target (Olsen P Index 2) and higher levels, responses are usually negligible
(Paynter and Dampney, 1991; Power et al., 2005). Soil type has an important influence on the yield
response to P fertiliser and the amount of phosphate fertiliser needed to build soil reserves from
one Index to another (Bolland et al., 2003). Phosphorus uptake and yield at any given P Index can
also be increased by improving soil structure, and thereby affect root growth and distribution, soil
aeration and access to water and nutrients (Ball et al., 2005).
Soils vary in their capacity to fix P and this influences the amount and frequency of phosphate that
needs to be applied to maintain soils at target P Index. Soil testing every 3-5 years allows soil P
reserves to be verified, but nevertheless in high P-fixing soils this could result in the amount of
extractable soil P declining to below the level that is critical for optimal production between
sampling dates. Consequently, in some countries, adjustments have been made to P
recommendations to account for the level of P-fixing in different soil types. For example, in
Scotland non-calcareous mineral soils have been mapped at soil association level as Index 1, 2
24
and 3 to reflect inherent soil phosphorus sorption capacity (PSC; Sinclair et al., 2015a). For
established grass/clover swards the target soil P status on PSC 1 and 2 soils has been lowered to
the lower band of moderate (M-; 4.5-9.4 mg P/l using the Modified Morgan method), but remains at
M+ (9.5-13.4 mg P/l) on PSC3 soils (Table 5). For grass only swards the target soil P status
remains at M- for all soils.
Research carried out in N.I. (Daly et al., 2015) supports the use of P sorption capacity to adjust P
recommendations in soils with different parent materials and chemical properties (e.g. calcareous
soils, extractable aluminium and soil pH).
Table 5. Effect of phosphorus (P) sorption capacity (PSC) on adjustments (kg P2O5/ha/year) to buildup or run-down soil P status for cereal-based arable rotations and established grass/clover swards
(source: Sinclair et al., 2015a - SRUC Technical Note TN668).
P sorption
capacity
Soil P status
Very low (VL)
Low (L)
Mod (M-)
Mod (M+)
High (H)
PSC1
+40
+20
0
-10
-20
PSC2
+60
+30
0
-20
-30
PSC3
+80
+40
+20
0
-40
Researchers in the Republic of Ireland (ROI) and N.I. have investigated the benefit of adjusting P
fertiliser recommendations according to relatively small changes in soil P status. Bailey et al.
(2014) summarised the evidence used to justify splitting the Olsen-P Index 2 range into a 2- Pbuilding range and a new 2+ target range for grassland in Northern Ireland. Using data from 12
farms in N.I., they found that at 2nd cut, Diagnosis and Recommendation Integrated System (DRIS)
P Indices, which provide a reliable measure of herbage P sufficiency status (Bailey et al., 1997),
were significantly and positively correlated with Olsen P, and declined to negative (deficient) values
when Olsen P fell below 20 mg P l-1 (i.e. the mid-point of the RB209 P Index 2 range).
Furthermore, at Olsen P Index 2-, swards had ‘Low’ herbage P status when RB209 P
recommendations were closely followed, but herbage P status was ‘Adequate’ when 15 kg P/ha
more than the RB209 recommendation was applied (P<0.05). Bailey et al. (2014) suggested that
the full P Index 2 range (16-25 mg P/l) should be split into equal sub-ranges, 2- and 2+, and that
proportionately higher P recommendations should be assigned to the 2-sub-range.
The NI soil and grass analysis results align with research carried out in the Republic of Ireland. For
example, Schulte (2007) outlines why ROI has opted for a higher target Morgan soil P test Index
for grassland; the new P-building range, i.e. Morgan Index 2 for Irish grassland, is equivalent to 1620 mg Olsen P/l, i.e. the new Index 2- ‘P-building’ range for NI grassland; and the new ‘Target’ P
25
range, Morgan Index 3 for Irish grassland is equivalent to the new ‘Target’ Index 2+ range for NI
grassland (although the Morgan Index 3 range extends to the equivalent of 30 mg Olsen P/l).
The amount of phosphate and potash to apply as ‘maintenance’ dressings is a function of the P
and K content of the grass sward. It is important therefore that the offtake estimates in
recommendations reflect the P and K contents in grass and clover at each cut or grazing through
the season. A number of research projects have investigated the P and K content of grassland
swards and how this varies with species composition, nutrient addition and time of year.
Høgh-Jensen and Schjoerring (2010) investigated variations in above-ground accumulation of N, P
and K in white clover and ryegrass grown separately or in mixture under field conditions in
Denmark (18 km west of Denmark) over three cuts (2 June, 7 July, 28 August, and 29 October) in
1998 and four cuts (31 May, 8 July and 12 October) in 1999. They reported that the interaction
between grass and clover grown together is complex and involves competition, facilitation and
complementarity. For instance, ryegrass grown with clover accumulated significantly more N, P
and K than when grown alone, conversely the growth of white clover was significantly reduced by
N application particularly when P and K were also applied. The proportions of P and K in ryegrass
shoots grown in monoculture were only half that in mixture with clover (Table 6). Conversely,
clover grown with ryegrass had a lower shoot P concentration compared to growing alone. Crush
et al. (2015) demonstrated that modern white-clover cultivars use soil-P more effectively compared
to older cultivars. Increased P-efficiency by white-clover, was due to repeated selection of shoot
traits, as there were no differences in root morphology or architecture between cultivars.
In Defra project IF01121, at twelve N response sites in England and Wales, phosphate and potash
concentrations in herbage were measured at each of four cuts from plots given the RB209
recommended rate of N, phosphate and potash and adjusted to 15-20% DM (Figure 6 and 7).
Mean phosphate offtakes were very similar to the offtakes values in the “Fertiliser Manual”.
However, potash offtakes at second, third and fourth cuts were 42-63% higher than the offtake
values in the manual. This has implications for potash fertiliser recommendations (although more
evidence is needed) and emphasises the importance of regular soil sampling and analysis to check
soil pH and nutrient reserves.
26
Table 6. Phosphorus (P) and potassium (K) offtakes (kg ha-1) of mixtures and pure stands of white-clover and perennial ryegrass (1 year
and 2 years after establishment) following the application of nitrogen (N), P, K fertiliser at different rates, applied during the first week of
April in each year. Taken from Høgh-Jensen and Schjoerring (2010).
Year
1
2
N, P, K rate
(kg/ha/yr)
Pure clover
Phosphorus offtake (kg/ha)
Grass
Clover mix
mix
a
13.3
7.7cd
Pure grass
Pure clover
4.3b
104d
Potassium offtake (kg/ha)
Grass
Clover mix
mix
c
69
62de
Pure grass
0, 0, 0
19.6ab
0, 0, 120
18.0b
8.9bc
7.1d
4.4b
173abc
106b
58e
29c
0, 20, 0
23.5a
12.7a
9.2cd
3.9b
148c
109b
70de
26c
0, 20, 120
22.0ab
13.4a
10.9c
4.5b
183ab
136a
96cd
30c
120, 0, 0
18.7ab
9.7b
15.0b
13.7a
101d
58c
128c
115b
ab
cd
b
a
ab
c
b
141a
6.4
120, 0, 120
19.0
120, 20, 0
24.0ab
7.6bcd
22.5a
12.9a
168bc
64c
181ab
117b
120, 20, 120
22.4ab
6.0d
21.8a
14.6a
199a
59c
216a
146a
0, 0, 0
13.9d
5.4c
7.8c
2.5c
79f
31d
57e
14c
0, 0, 120
16.4c
9.5b
9.9c
3.3c
144bc
99b
82d
20c
0, 20, 0
19.5b
12.2a
14.6b
3.7c
126de
66c
101cd
22c
0, 20, 120
21.9a
13.9a
13.6b
2.9c
141cd
124a
98d
18c
120, 0, 0
14.9cd
5.6c
12.9b
11.3b
74f
29d
100cd
88b
c
b
b
ab
c
b
132a
8.3
17.0
14.0
13.6
13.2
176
16.4
120, 20, 0
21.0ab
8.9b
18.3a
14.5a
123b
42d
120c
124a
120, 20, 120
19.7ab
8.9b
18.8a
14.6a
173a
70c
176a
129a
27
74
175
120, 0, 120
Values followed by different letters are significantly different (P <0.05)
157
b
67
23c
146
2.0
1.8
P2O5 kg/t FW
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Cut 1
Cut 2
Cut 3
Cut 4
Figure 6. Mean phosphate (P2O5) offtake values at each of four cuts at the twelve IF01121 N response
sites. Error bars represent one standard deviation from the mean. The hatched line indicates the
phosphate offtake value for fresh grass (15-20% dry matter) in the “Fertiliser Manual” of 1.4 kg P2O5/t
fresh material.
10
9
8
K2O kg/t FW
7
6
5
4
3
2
1
0
Cut 1
Cut 2
Cut 3
Cut 4
Figure 7. Mean potash (K2O) offtake values at each of four cuts at the twelve IF01121 N response
sites. Error bars represent one standard deviation from the mean. The hatched line indicates the
potash offtake value for fresh grass (15-20% dry matter) in the “Fertiliser Manual” of 4.8 kg K2O/t
fresh material.
Higgins et al. (2012) (Section 4.1) reported that lime applied annually (studied over 3 years) to a
permanent grassland has a cumulative effect and by the third year of application, significantly
28
increased the P-removal of the herbage (30.7 kg P2O5/ha) compared to treatments receiving no
lime (28.6 kg P2O5/ha). The authors suggest that the lime may have increased the mineralisation of
soil P or stimulated root growth.
The effects of different potash fertiliser rates on forage grass yields, K herbage content and soil
fertility were assessed in demonstration plots, on a sandy loam soil in Warwickshire (Potash
Development Association (PDA) leaflet 5b, 2007a). At the start of the trial soil was at K2O Index 1
and K2O was applied to plots for 4 consecutive years at three different rates: 0, 160 and 320 kg
K2O/ha. In year one, the site responded to potash with yield increases of 12-15 %, and by years 3
and 4, yields on the 320 kg K2O/ha were double that relative to the control. Overall, large quantities
of potash were removed in the silage and it was found that the application of 160 kg K2O/ha did not
maintain the initial soil K2O levels despite yields and K offtakes being lower compared to the higher
K2O rate. Applications of 320 kg K2O/ha helped to maximise silage yields, K offtakes and also
maintained or slightly improved soil potash levels. Silage analysis demonstrated that at maximum
yields, N and K are removed at similar amounts, indicated by N:K ratios of 1:1 therefore, both N
and K should be replaced by similar amounts to maximise yields and nutrient efficiency.
The K2O requirements of perennial ryegrass/white clover were assessed in a 3-year field trial
undertaken by Kingshay Farming Trust (PDA leaflet 26, 2007). In this study, the mean potash
content of perennial ryegrass/white clover silage at 25% DM was 6.8 kg K2O/t of fresh material
(averaging 6.7 kg K2O/t at first cut; 7.0 kg K2O/t at second cut; and 6.4 kg K2O/t at third cut), which
was consistently higher than the RB209 8th edition standard value of 6.0 kg K2O/t.
Finally, there is some evidence that changes in nutrient management practices may have
increased P use efficiency in some European countries. For example, Milhailescu et al. (2015)
compared P use efficiency (PUE) and farm-gate P balances on intensive grass-based dairy farms
in Ireland before and after the implementation of good agricultural practice regulations (GAP).
Comparisons indicate that post GAP introduction there was a reduction in P surpluses by 74% per
hectare and 81% per kg of milk solids. The improvements in P balance were due to both a
reduction in P fertiliser application rates and improvements in P management, including a shift to
spring application of organic manures.
4.3.1.
Conclusions for grassland phosphate and potash recommendations
Recommendations have been developed in Scotland to take account of the P-fixing capacity of
different soils, based on a map of soil parent material types (as defined by soil associations) and
the contrasting P response of different crops. A similar approach could be developed for England
and Wales by assessing the P-fixing capacity of different soil associations.
29
Research in ROI and NI indicates that Olsen P Index 2 could be split into a lower and upper subband with grass P recommendations increased for the lower sub-band. Similar research is required
in England and Wales to determine the herbage P sufficiency status at different levels of Olsen P
and P fertiliser use.
Data from Defra project IF01121 and the PDA (2007) indicates that the fresh grass potash offtake
values in the “Fertiliser Manual (RB209)” may underestimate the amount of potash removed within
cut grass systems. However, the IF01121 and PDA data is limited in amount (14 sites). More data
is needed for which the grassland management (primarily cutting date, K Index and nutrient
applications) is known.
4.4.
Response of grass to lime
Higgins et al. (2012) in a field study in N.I., assessed the effect of annual applications of pelletized
dolomitic lime compared to ground lime on grass (sward predominantly meadow grass and
ryegrass) productivity and soil chemical properties when receiving N-fertiliser applications (at 0, 75,
150, 225 and 300 kg N/ha) as calcium ammonium nitrate. Overall, the study found that there was
no difference in any of the parameters measured (including dry matter yield; Table 3) when
applying pelletized lime compared to ground lime.
4.4.1.
Conclusions for liming recommendations
It is clear from Higgins et al. (2012) that the ability of different liming products to raise pH by a
certain amount on any given soil is determined by application rate and neutralising value (NV). The
speed of change in pH is determined by the fineness and hardness of the liming product and
associated solubility. There is therefore no need to change the recommendations. However, the
RB209 Technical Working Group felt that inclusion of the liming factor in the RB209 lime
recommendations table would aid the clarity and precision of the recommendations. Consideration
should also be given to the inclusion of advice on the use of seashell sand and its impact on soil
pH.
Sinclair et al. (2014) provide advice on GPS sampling for soil pH and the variable rate application
of lime. These recommendations will be incorporated into the “Nutrient Management Guide
(RB209)”.
4.5.
Impact of grazing upon herbage production
The DM yield response of grassland to applied nutrients varies according to grassland
management in terms of the frequency of defoliation, whether the sward is cut or grazed and the
level of N fertiliser use. It is acknowledged that the DM yield ranking of grass cultivars can vary
30
according to whether they are tested under long-cycle (infrequent cutting), short-cycle (simulated
grazing) or livestock grazing conditions (Jafari et al., 2003). Indeed, cultivars that perform better
under a cutting regime may be inferior under animal grazing. This is supported by observed
differences in sward structure between cut and grazed swards (Smith et al., 1971). Grazed swards
tend to produce lower dry matter yields than less frequently defoliated cut swards (Binnie and
Chestnutt, 1991; Boswell, 1977; Wilkins, 1989; Cashman et al., 2016). This is due to a number of
factors including higher overall Green Area Index (GAI) and higher DM yield per tiller under the
less frequent defoliation within a silage cutting regime (Matthew et al., 1996), as well as the effects
of treading, plant pulling and selective grazing, particularly at higher N fertiliser rates (Evans et al.,
1998; Richards, 1978).
It is unclear if the difference in sward heights resulting from selective grazing can affect annual DM
yields when comparing cut and grazed swards managed at the same average height and
defoliation frequency. DM yields of perennial ryegrass under a non-uniform defoliation regime
(simulating grazing) were similar to those obtained under a uniform defoliation regime (simulating
cutting), although the frequency of defoliation and the amount of fertiliser applied may influence the
outcome; Smith et al. (1975) and Remison and Snaydon (1980) in Dale et al., (2013). Dry matter
yield and digestibility under simulated grazing has been shown to correlate well with yields under
rotational livestock grazing, particularly in established swards (Cashman et al., 2016).
Reduced dry matter offtakes at higher defoliation frequencies is supported by data collated from
Reaseheath College, Rothamsted Research North Wyke and the AHDB ‘Low N Grass’ project
(74316). Since the replicate yields were not available, the mean DM yields were plotted rather than
fitting a specific growth curve (Figure 8 and Figure 9). All the sites had high SNS status with mean
y0 (i.e. yield at 0 N kg/ha) at Reaseheath in 2010-12 (7 or 8 cuts) measured at 6 to 7 t DM/ha and
mean y400 at 8.8 to 11.5 t DM/ha. The sites where 5 to 6 cuts were taken yielded similar quantities
of grass DM as the moderately yielding IF01121 sites (Figure 1) with y0 (DM yield at 0 kg N/ha) at
4.5 to 6.3 t DM/ha and y400 at 11.6 to 12.8 t/ha.
31
5‐6 cuts
20
18
Yield (t/ha dm)
16
14
12
10
8
NW 2011 (6)
6
NW 2010 (6)
4
Low N Grass ‐ PRG 2013 (5)
2
0
0
100
200
300
N applied (kg/ha)
400
500
600
Figure 8. Grass dry matter (DM) yield response to manufactured nitrogen (N) fertiliser from field
experiments carried out in the UK where 5-6 cuts of grass were taken and 3 to 6 levels of fertiliser N
were used up to at least 400 kg N/ha. The legend indicates the site reference (and the number of cuts
taken).
At the ‘Low N Grass’ sites in 2013-14, where 9 cuts were taken, mean y100 was 3.8 to 4.5 t DM/ha
and mean y400 was 9.5 to 10.9 t DM/ha. This contrasts with y400 values at the IF01121 sites that
typically ranged from 12.5 to 16.9 t DM/ha (Figure 1). The differences in yield between the 2010-14
experiments using 8-9 cuts of grass and the IF01121 field experiments (2012-14 using 4 cuts)
confirm that data from a three or four cut system should not be used to predict typical yields from
multiple cut or livestock grazing systems; and vice versa (Cashman et al., 2016). This may partly
explain the difference in DM yields between the IF01121 experiments and the N response trials
from the 1970s and 80s (such as GM 20, GM 23 and GM24) used to produce indicative yields for
the 8th edition of RB209. For example, the GM23 series of 22 trials (1978-79) used 6 cuts of grass
(Morrison, 1980; 1987).
The DM yield N response curves used in the “Fertiliser Manual (RB209)” were derived from a
selection of datasets from the GM20, GM24 (cutting data only) and GF01 (permanent pasture
treatments only) trials carried out from 1970 to 1993, amounting to 148 site years. The number of
datasets in Good/Very good, Average and Poor/Very poor Grass Growth Classes (GGC) were 44,
88 and 16 site years, respectively. The derived N response curves were used to predict the
amount of N fertiliser to apply for both cutting and grazing land. Given what is known about the
contrasting production potential of cut and grazed swards and the apparent improved N use
32
efficiency and productivity of modern perennial ryegrass varieties, it is likely that the N response
curves used within the 8th edition model approximately reflect the production potential of modern
grazed swards, but underestimate the production potential of cut swards. This has implications for
the grassland N recommendations themselves, with the effect from this consideration alone being
a reduction in the N fertiliser required to meet a given level of energy requirement. However, it is
also important to take account of the assumptions used to calculate the effect of other key factors
within the energy model, such as livestock energy requirements themselves, grass utilisation
efficiency (spoilage) and concentrate substitution effects (the suppression of grass/silage DM
intake per kg of concentrate DM fed), since these factors also have important and sometimes
opposing effects on the N fertiliser recommendation (Dale et al., 2013).
7‐9 cuts
Reaseheath 2010 (7)
Reaseheath 2012 (7)
Reaseheath 2011 (8)
Low N Grass ‐ PRG 2013 (9)
Low N Grass ‐ PRG 2014 (9)
20
18
Yield (t/ha dm)
16
14
12
10
8
6
4
2
0
0
50
100
150
200
250
300
350
400
N applied (kg/ha)
Figure 9. Grass dry matter (DM) yield response to manufactured nitrogen (N) fertiliser from field
experiments carried out in the UK where 7-9 cuts of grass were taken and 3 to 5 rates of fertiliser N
were used up to 400 kg N/ha. The legend indicates the site reference (and the number of cuts taken).
4.5.1.
Grazing height
Pre-grazing herbage mass can have a significant impact on grass dry matter intake (GDMI),
pasture quality and milk production; current guidance recommends that grass covers are kept low
in the spring. For instance, studies have shown that lower pre-grazing herbage mass (from 7950 to
2180 kg DM/ha) increased GDMI due to, higher leaf proportions and lower stem and dead material
proportions (Hodgson and Wilkinson, 1968, in Wims et al., 2014). Furthermore, lowering spring
herbage mass can increase overall milk production (e.g. Curren et al. 2010).
33
Wims et al. (2014) in a field experiment located in Ireland, compared the effect of 3 different pregrazing herbage mass treatments (Low: 1,150; medium 1,400 and high: 2,000 kg DM/ha) on
perennial ryegrass pasture and dairy cow productivity. In contrast to current guidance, it was found
that grass production was significantly (P<0.01) reduced on the low-treatment (10,142 kg DM/ha)
compared to the high-pre-grazing herbage treatment (12,112 kg DM/ha). Notably, cows required
more grass silage supplementation (+73 kg DM/ha) during the grazing season, demonstrating that
there can be an increased requirement for purchasing feed when maintaining very low herbage
mass.
Phelan et al. (2013a) recommended that for grass-clover mixes and optimal herbage production,
swards should be grazed to 4 cm height; while there was no evidence that post-grazing height (4, 5
or 6 cm) had an effect on the white-clover content. In contrast, a further study, on a grass-clover
sward in Ireland, Phelan et al. (2014) found that reducing the grazing height from 6 to 2.7 cm (in
the summer to winter period) increased both clover content and clover yield in the sward.
Furthermore, a defoliation interval of 42 days achieved the highest total yield at 11 t DM/ha.
Overall, Phelan et al. (2014) recommended a defoliation interval of 42 days combined with a
grazing height of 2.7-3.5 cm for grass-clover swards.
4.5.2.
Treading by livestock
Treading by cattle is another factor associated with grazing that can impact on herbage yields.
Glasshouse experiments have indicated that white-clover on ‘wet’ soils can be more susceptible to
damage by treading compared to perennial ryegrass (e.g. Grant et al. 1991). In a field experiment,
Phelan et al. (2013b) assessed the effect of treading on clover content, herbage production and
soil properties within three clover based grazing systems on a ‘wet’ soil in Ireland. The study found
that treading reduced (P <0.001) annual yields of white clover and perennial ryegrass by similar
amounts, with yield reductions of 0.45 and 0.59 t/ha, respectively. In contrast to earlier studies, it
was concluded that there was no difference in the susceptibility of white-clover or ryegrass to
damage by treading.
Frame and Laidlaw (2014) reported that soils with a high organic matter content and associated
higher water retention capacity are particularly susceptible to poaching. They found that the risk of
damage was also highest on re-seeded swards that have not yet developed sufficient root mass to
strengthen the soil surface.
4.5.3.
Grazing returns
On grazed systems it is important to account for the nutrients recycled from dung and urine,
although the uneven nature of livestock excreta deposition creates problems for predicting its
34
influence on DM yield response at the field scale. It is estimated that >70% of the consumed N is
recycled through the direct decomposition of animal excreta and that inorganic soil N under urine
patches can be up to 10 times greater than under dung patches and more than 30 times greater
compared to areas free from animal excreta (Afzal and Adams, 1992, in Eriksen et al. (2015)).
However, only a proportion of a grazed field will receive nutrients from livestock excreta in any
given season. Overlapping of excreta patches tend to occur in parts of the pasture where animals
gather. For example, Betteridge et al. (2010) monitored grazing steers on a steep 11 ha hill
paddock and found that cows camped on flatter areas, with only 10% of the paddock containing
61% of all urine patches excreted during the grazing period (12 days).
The proportion of pasture covered by urine and dung patches depends on estimates of the mean
size of urine or dung patches. For example, the proportion of pasture covered by urine patches
over an annual grazing season, at a stocking rate of 3.2 cows/ha, increases from 23%, when
considering a urine patch of 0.33 m2, to 33% for a larger urine patch of 0.5 m2 (Dennis et al., 2011,
in Dale et al., 2013). On a New Zealand dairy farm with a stocking rate of 3 cows/ha/yr, Haynes
and Williams (1993) calculated that 23% of the pasture would be covered in excreta (urine and
faeces) after one year.
Dale et al. (2013) estimated that, even at relatively high stocking rates of 3 cows/ha, up to 30-50%
of the pasture area is affected by urine and dung patches every year (Richards and Wolton, 1976
etc.), and will benefit from enhanced herbage growth following nutrient returns from animal excreta.
Excretal N recovery by herbage is low, being on average 27% for urinary N and 11% for faecal N
(averages from numerous studies summarised in Dale et al., 2013). However, several studies
indicate that a lower recovery of urinary N occurs when the sward receives inorganic fertiliser N in
addition to urine and faeces applications (e.g. Deenen and Middelkoop, 1992; Di et al., 2002).
Measurements of N leaching losses under different management systems can provide an
indication of the amount of N recycled at grazing. Eriksen et al. (2015) investigated the effect of
management practices (i.e. combinations of cutting, grazing and spring slurry application) on
nitrate leaching from a grass-clover ley site in Denmark. It was found that a combination of spring
slurry application (100 or 200 kg total N/ha) and grazing resulted in the highest N leaching losses
of c. 60 kg N/ha. However, when either grazing or spring slurry application was carried out alone
nitrate leaching losses were reduced to c. 20 kg N/ha.
A dairy cow urinates on average 9 times per day (Aland et al., 2002; White et al., 2001) and
defecates 12 times per day on average (Aland et al., 2002; Haynes and Williams, 1993; Orr et al.,
2012; White et al., 2001). If we assume from this that 40% of excreted N is derived from urine and
60% from dung, and that 27% of excretal N is recovered from urinary N and 11% from faecal N,
35
the average N recovery is approximately 17%. Assuming lower N recovery when swards receive
inorganic fertiliser N in addition to urine and faeces applications (Deenen and Middelkoop, 1992; Di
et al., 2002), this can be rounded down to 15%. We therefore conclude that at low to moderate N
rates (often corresponding to low or moderate stocking rates), grazing has a positive effect due to
approximately 15% recovery of applied N. This positive grazing effect results in higher grass yields
under livestock grazing than under simulated grazing (i.e. short-cycle cutting). This effect is likely to
occur at N rates up to 300 kg N/ha (Richards, 1978; Richards and Wolton, 1976), but it is difficult to
predict as it depends on local environmental conditions and seasonal effects.
In the 8th edition of RB209 no account was taken of nutrient returns at grazing. By contrast, in the
7th edition the N recommendations for grazing dairy cattle on Average, Poor or Very Poor GGC
land were reduced by 40 kg N/ha relative to the cutting recommendations to take account of
grazing returns (Dampney, 1992). However, for grazing on Good and Very good GGC land, N
rates were increased by 40 kg N/ha and 80 kg N/ha respectively, compared with rates under a
cutting regime.
4.5.4.

Conclusions for grazing and grassland management recommendations
Nutrient cycling at grazing is complex. Nutrient returns from excreta is uneven, particularly
for N in urine patches. Nevertheless, the research indicates that around 15% of excretal N
is typically recovered by a grass sward and this figure has been used to adjust N
recommendations at grazing. The uneven deposition of phosphate and potash by grazing
livestock suggests that there could be clear advantages from GPS sampling for soil pH and
soil nutrient reserves in grassland fields. This advice will be incorporated into the section on
the “Principles of grassland nutrient management”.

DM yields tend to be lower over shorter grazing cycles compared with longer cutting cycles.
Research in ROI indicates that cutting regimes can produce DM yields that are c.20%
higher on average than under a grazing regime (Cashman et al., 2016). This has
implications for the amount of fertiliser N (and other nutrients) to apply to cutting and
grazing land to achieve target grass DM yield and quality. However, there is currently not
sufficient data on grass DM yields under grazing management in England and Wales. The
approach in the 8th edition of RB209 that used the same N response curves for cutting and
grazing has therefore been retained. Indeed, on 26th April 2016 the Livestock Technical
Working Group (TWG) agreed that the GM20/23/24 data should continue to be used to
underpin the recommendations with indicative DM yield ranges used to illustrate the
potential of modern grass varieties. However, adjustments are made for nutrient returns at
grazing (see above).

Grazing height can influence sward recovery, DM yield and overall productivity of grazing
livestock systems. However, there are a number of approaches within the industry and
36
advice on herbage mass at the start and end of each grazing cycle should not form part of
nutrient management recommendations.

Treading and trampling by livestock can have a negative impact on sward productivity.
Nutrient management guidelines should therefore signpost users to industry advice on
grassland and soil management related to the risks of causing compaction or poaching.
4.6.
4.6.1.
Grass and silage quality
Effect of nitrogen rate
Factors which can influence silage quality include: grass species, rate of N application and stage of
maturity at harvest. Bednarek et al. (2015) assessed the effect of mineral NPK fertiliser
applications on N-fractions (i.e. total, protein, mineral, ammonium and nitrate nitrogen) within
Timothy grass in a three year replicated field experiment in Poland. Nitrogen fertiliser was applied
as ammonium nitrate (AN) at 3 rates: 120, 240 and 360 kg N/ha. It was found that the content of
total, protein, ammonium and nitrate nitrogen was positively correlated with the rate of mineral
fertiliser application, mainly N and phosphate and to a lesser extent potash application rate.
However, even at the higher N fertiliser rate, the Timothy grass did not contain excessive amounts
of ammonia or nitrates and was a valuable bulk feed, indicating that whole season N fertiliser rates
up to 360 kg N/ha did not have a detrimental effect on grass silage quality.
King et al. (2013) investigated the impacts of N fertiliser rate (0 or 125 kg N/ha) and harvest date
on silage quality (fermentation characteristics and aerobic stability) and dry matter yield of silage
produced from 5 different grass species; perennial ryegrass (Lolium perenne cv. Gandalf), italian
ryegrass (Lolium multiflorum cv. Prospect), tall fescue (Festuca arundinacea cv. Fuego), cocksfoot
(orchardgrass, Dactylis glomerata cv. Pizza) and timothy (Phleum pratense cv. Erecta). Five
harvest dates were distributed fortnightly from 12 May to 7 July. Overall, it was found that there
was little effect of N fertiliser on the extent or direction of fermentation with the ryegrass and tall
fescue silages, which exhibited a lactic acid dominant fermentation. However, during fermentation,
timothy and cocksfoot had higher pH (>4.2), butyric acid (> 10 g/kg dry matter) and ammonia-N
(>100 g/kg total N) levels, indicative of secondary clostridial activity during storage; and the italian
ryegrass herbage incurred the greatest dry matter losses during ensiling, particularly following the
early harvest dates, suggesting yeast fermentation of sugars.
Durant & Kerneis (2015) assessed the effect of manufactured fertiliser and organic manure
applications on forage yield and feed value (crude protein content and digestibility) in a 7-year
experiment located on a permanent grassland in Western France. The results demonstrated that N
applied at rates of 60 and 100 kg N/ha/yr did not improve feed value (crude protein content and
digestability), however, after the first year of the experiment it did improve DM yield.
37
As part of the Defra IF01121 validation project, herbage quality was measured on each plot at
each of four cuts at twelve sites in England and Wales. There was a relationship (P<0.05) between
fertiliser N rate at first cut and herbage quality at four out of the ten IF01121 sites. There was a
significant correlation for crude protein (CP) and metabolisable energy (ME) at 3 out of 10 sites,
and for neutral detergent fibre (NDF) at 2 out of 10 sites. However, at only two sites (both in
Cheshire) was there a relationship between fertiliser N rate and all three key measures of herbage
quality.
Crude protein (CP) is a measure of the nitrogen content of the cut grass and indicates the maturity
of grass at the time of cutting, and the time interval between N fertiliser application and cutting. A
CP content of 120-150 g/kg (12-15%) indicates that grass was cut at an optimal stage of growth
(Corrall et al., 1982). At a long term grass site in Devon, CP ranged from 11.3 to 19.0% with ‘target’
concentrations measured at 40-80 kg N/ha (Figure 10).
250
P=0.009
c
Crude protein (g/kg)
200
bc
bc
ab
bc
150
a
100
50
0
0
40
80
120
N applied (kg/ha)
140
180
Figure 10. The relationship between fertiliser nitrogen (N) rate and crude protein for first cut silage at
a long term grass site in Devon. Values with the same letter are not significantly different (P >0.05).
Error bars represent the standard error of the mean.
High protein concentrations can indicate high ammonia levels, which can result in poor
fermentation and waste in the silage clamp. High ammonia levels can be caused by high protein
levels in grass at cutting or by cutting young, low sugar content grasses when wet. It is an
important consideration when planning fertiliser N application rates (Thomas et al., 1991).
38
Metabolisable energy (ME) is a measure of the energy value of silage expressed as the amount of
energy contained in every kg of grass/silage dry matter. The younger and drier the grass, the more
energy the grass/silage will supply for milk and liveweight gain (Yan et al., 1997). There was no
relationship between ME and N fertiliser rate at seven out of ten sites. However, at three sites, ME
decreased with increasing fertiliser N rate (e.g. Figure 11), with energy value reducing from ‘top
quality’ (11.4-11.7 ME) at 0-40 kg N/ha to ‘average/good’ (10.1-10.8 ME) at 180 kg N/ha.
12
10
ME (MJ/kg)
8
P=0.01
6
4
2
0
0
40
80
120
N applied (kg/ha)
140
180
Figure 11. The relationship between fertiliser nitrogen (N) rate and Metabolisable energy (ME) for first
cut silage at a long term grass site in Cheshire. Error bars represent the standard error of the mean.
The straight line indicates the ME value (11.5 MJ/kg DM) used in the energy model underpinning the
8th edition grassland recommendations.
ME values were generally ‘good’ (10.6 to 11.6 ME) at N fertiliser rates between 40 and 140 kg
N/ha. There was therefore some indication that ME values can reduce below the levels assumed in
the 8th edition energy model at higher rates of fertiliser N. A reduction in ME of 1 MJ/kg DM (i.e. an
ME value of 10.5 MJ/kg compared with 11.5 MJ/kg used in the energy model) would have the
effect of increasing the total N fertiliser requirement by c. 20 kg N/ha to achieve the same level of
ME per hectare (i.e. an additional 20 kg N/ha would be needed to achieve the same amount of
ME/ha, particularly for first cut fertiliser N rates above 100 kg N/ha at which ME can be
suppressed). However, this effect was only measured at three out of ten sites and it was not
possible to predict when it occurs.
Neutral detergent fibre (NDF) is a measure of the total fibre in grass/silage and also indicates the
bulkiness of the feed and the likely level of intake (Corral et al., 1982; Minson, 1990). Young grass
silage tends to have an NDF of 45–50%, and mature grass silage 60-65%. Late cut, ‘stemmy’
39
silages have the highest NDF values (Corral et al., 1982). High NDF values tend to indicate lower
digestibility, energy and protein values, but NDF can improve intake and rumen health (Minson,
1990). The ideal NDF range in grass/silage is 50-55% (500 – 550 g/kg DM; Corral et al., 1982).
There was a positive relationship between fertiliser N rate and NDF at two out of ten sites (e.g.
Figure 12). This was supported by findings from the AHDB ‘Low N Grass’ project (74316).
600
P<0.001
cd
500
d
cd
140
180
bc
NDF (g/kg)
400
a
ab
300
200
100
0
0
40
80
120
N applied (kg/ha)
Figure 12. The relationship between fertiliser nitrogen (N) rate and neutral detergent fibre (NDF) for
first cut silage from a third year ley in Cheshire. Values with the same letter are not significantly
different (P >0.05). Error bars represent the standard error of the mean.
Overall, fertiliser N rate did not have an important effect on herbage quality, with no relationship
between N rate and key measures of herbage quality at most sites. Assuming moderate to high
output systems, at the few sites where there was a reduction in ME values with increasing N rate it
could still be in the farmer’s interest to use higher N rates to produce more ME overall, although
lower ME in terms of MJ/kg DM could reduce milk yield per cow if not compensated for by
increased concentrate use. However, higher ammonia-N levels at the highest N rates (140-180 kg
N/ha) may result in poor fermentation in the clamp.
4.6.2.
Micronutrients
Deficiencies of micronutrients can result in major reductions in the health, fertility and productive
performance of livestock. Fifteen micronutrient elements are believed to be essential for animal life:
iron, iodine, zinc, copper, manganese, cobalt, molybdenum, selenium, chromium, tin, vanadium,
fluorine, silicon, nickel and arsenic (Suttle, 2010). The availability of many of these elements, such
as cobalt, copper and selenium, does not restrict grass growth, but too little in the overall diet can
40
lead to deficiency in some animals; and cobalt deficiency reduces Vitamin B12 production in
rhizobium, thereby lowering N fixation and associated productivity (Sinclair et al., 2015b). The aim
should be to only use micronutrient supplementation where deficiency has been diagnosed.
Selenium (Se) is an essential trace element for animal health. Borowska et al. (2011) investigated
the effect of applying farmyard manure (FYM) manure at different rates (0, 20, 40, 60 and 80 t/ha)
on Se contents in the soil and in red clover. Applying manure at 80 t/ha resulted in a 1.7 fold
increase in the Se concentration in the soil (mean background Se content = 0.101 mg/kg).
Furthermore, the highest selenium concentrations in red clover were measured from plots
receiving 40 and 60 t FYM/ha; about 25% higher than the control.
Sinclair et al. (2015b) provide advice on the management of cobalt (Co) in grassland soils. Soils in
Scotland have been mapped as “high”, “moderate” or “low” risk for Co deficiency, according to
parent material, soil drainage characteristics and soil texture. High risk soils include organic soils;
drifts derived from acid schists, granulites, granitic rocks, greywackes and shales; and fluvioglacial
sands and gravels. Interpretative scales for soil extractable Co concentrations (mg/kg) are also
provided and soil testing is advised in areas mapped as “high” predicted risk; and on peaty soils
herbage analysis is also advised. Co deficiency can be corrected for approximately 4 years
through the application of hydrated cobalt sulphate to pasture in early spring. However, there are
concerns over the use of Co salts as suspected carcinogens, and suppliers of trace elements for
livestock in the UK prefer to provide intraruminal boluses or drenches containing cobalt to
supplement grass. Selenium (Se) can also be supplied as a grassland fertiliser in deficiency
situations (Scott, 2010). However, feed and forage supply needs to be carefully integrated, since
for some micronutrients, such as Se, the difference between deficiency and toxicity can cover a
narrow range of concentrations in the diet.
4.6.3.
Effect of Biostimulants
At the Experimental Unit of the University of Natural Sciences and Humanities in Siedlce (Poland),
Godlewska and Ciepiela (2015) investigated the effect of the biostimulant Kelpak® (a seaweed
extract containing plant growth regulators cytokinins and auxins and amino acids) on the content of
micronutrients in two grass species grown in monoculture: Dactylis glomerata L. (cv. Amila) and
Festulolium braunii (K.Richt.) A. Camus (cv. Felopa) grown in a monoculture.
Kelpak SL was applied at 0 (control) and 2 dm3/ha, and nitrogen was applied at 0, 50, 100 and 150
kg/ha, with grass cut three times in each year (2010-2012). The application of Kelpak consistently
and significantly increased Zn, Cu, Fe and Mn concentrations in the grass species tested,
regardless of other factors, although differences were numerically small (e.g. mean Zn content at
100 kg N/ha was 25.9 mg/kg DM with Kelpak SL and 23.8 mg/kg DM without; mean Cu content at
41
100 kg N/ha was 6.7 mg/kg DM with Kelpak SL and 6.0 mg/kg DM without). The concentrations of
Zn, Cu and Fe decreased with increasing N rate, while the Mn concentration increased. The
application of Kelpak increased the Fe:Mn ratio in the dry matter of both grasses, which implies an
increased risk of manganese deficiency (Malhi et al., 1998).
In two similar studies, Ciepiela et al. (2013) and Ciepiela and Godlewska (2015) found that Kelpak
in combination with N fertiliser (applied at 0, 50, 100 and 150 kg/ha) significantly increased the
yields and chlorophyll content of two perennial ryegrass cultivars and four grass-red clover
mixtures. The highest DM yields, protein and chlorophyll contents were measured when Kelpak
was sprayed and N fertiliser applied at a rate of 150 kg/ha. Kelpak had a small, but statistically
significant (P<0.05) additional effect to the N fertiliser.
4.6.4.

Conclusions for grass and silage quality recommendations
At some sites, application rates of fertiliser N above 120-140 kg N/ha for first cut silage
reduced the quality of cut grass in terms of ammonia-N content and ME. Although the stage
of maturity at harvest can have an influence, it is difficult to predict the agro-climatic or
growing conditions in which this deterioration in grass/silage will occur and therefore
manufactured N fertiliser rates above 120 kg N/ha should be avoided.

The importance of micronutrients for grass/clover productivity and animal health needs to
be made clear in the recommendations and options for correcting potential deficiencies
provided, including the use of micronutrients in fertilisers (e.g. selenium). The possible use
restriction of cobalt under the Registration, Evaluation, Authorisation and Restriction of
Chemicals (REACH) regulations should be borne in mind for the 2017 AHDB “Nutrient
Management Guide (RB209)”.

Studies in Poland indicate that biostimulants containing seaweed extract (from the kelp
species Ecklonia maxima) when applied with or without N fertiliser can improve grass DM
yield and quality in terms of protein and micronutrient content, although there were some
concerns over the biostimulant increasing the risk of manganese deficiency due to
increases in the Fe:Mg ratio. Research should be carried out to determine whether
biostimulants can increase yield and quality of grass/clover swards in England and Wales,
which have contrasting soils and agro-climatic conditions compared with eastern Poland.
42
5.
Grassland N recommendation systems
This section outlines the grassland N recommendation systems currently used within Scotland, the
Republic of Ireland and England and Wales, and compares the different approaches in relation to
the factors that are taken into account within each system.
5.1.
Fertiliser N recommendations in Scotland – SRUC Technical Note 652
The N recommendations for established grassland in Scotland (SRUC, 2013) involve the following
two stages:
i.
Select the ‘site class’ (equivalent to Grass Growth Classes in RB209); defined on a 1 to 5
scale based on rainfall and soil type (Table 7), with site class 5 having about half the grass
growth potential of site class 1.
ii. Identify the ‘appropriate’ nitrogen rates and sequences (kg N/ha) for each grass field based
on site class and grass management strategies, expressed as a set of ‘defoliation’
sequences (
iii. Table 8).
The technical note states that “in practice, levels of N use may be less than the figures shown…
(see Table 8)…to reflect the level of intensity and production that is required on that particular farm
unit”.
The nitrogen rates and sequences are derived from whole season “standard or maximum”
recommendations, which are reduced by 10kg N/ha with each site class, such that
recommendations for site class 5 are 40 kg N/ha less than for site class 1. When these differences
are distributed between multiple ‘defoliations’ the resultant differences between site classes are
relatively small
Table 7. Definition of site classes within the nitrogen recommendations in Scotland.
Soil texture
Average Apr-Sep rainfall (mm)*
More than 500
425-500
350-425
Less than 350
Site class **
Sands and shallow soils
2
3
4
5
All other soils
1
2
2
3
*Approx. 50% annual rainfall
**Add 1 for farms above 300 m
43
Table 8. A sub-section of the nitrogen (N) recommendations for established grassland showing
appropriate application rates and sequences for site classes 1 to 3 (SRUC, 2013).
Grass management
Defoliation
sequence
Site Class 1
Site Class 2
Site Class 3
120-90
120-90
120-90
SSG
120-90-60
120-90-60
120-90-60
SSS
120-90-70
120-90-70
120-90-70
120-90-70-30
120-90-60-30
120-90-60-20
120
120
120
120-70
120-70
120-70
120-60-50
120-60-50
120-60-50
120-60-50-40
120-60-50-40
110-60-50-40
90
90
90
80-60
80-60
80-60
80-60-50
80-60-50
80-60-50
80-60-50-40
80-60-50-40
80-60-50-40
SS
2 or 3 cuts silage + grazing
Nitrogen application rate (kg/ha)
SSSG
S
1 cut silage + grazing
SG
SGG
SGGG
G
Grazing with low clover
GG
GGG
GGGG
G = grazing; S = Silage
The Scotland N recommendations do not take account of Soil Nitrogen Supply.
5.2.
Fertiliser N recommendations in the Republic of Ireland (ROI)
The N recommendations for established grassland in Ireland (Coulter and Lalor, 2008) are split
between grazed and cut grass. The N fertiliser recommendations for grazed grass are essentially
based on stocking rate and involve the following four stages:
i.
Select the advised N application rate according to your stocking rate (Table 9). The advised
rates are “for swards ≥ 3 years old with no clover and of average soil–N fertility when
grazed by bovines”.
ii. For good ryegrass swards less than 3 years old, an additional 25% N may be applied
where necessary, provided that the rates do not exceed those prescribed in the ROI
Statutory Instrument (SI) 610 of 2010.
iii. Make adjustments for soils of lower than or greater than average fertility: “At stocking rates
below 200 kg/ha N, rates of N greater than those shown in this table can be applied on
poorer soils. Lower N rates may be appropriate on soils with above average natural
fertility”.
iv. Select the optimum timing of N applications and the appropriate annual start and end of N
fertiliser application according to the length of the grazing season (Table 10; Collins and
44
Cummins, 1996). “While these dates are recommended, decisions in each individual year
will need to be adjusted depending on the prevailing weather conditions”.
The recommendations include the note that “management should promote clover growth as a good
clover sward will reduce N requirements or make N fertiliser application unnecessary”.
Table 9. Available nitrogen (N) rates for swards that are greater than 3 years old, with no clover and
of average soil-N fertility when grazed by bovines (Coulter and Lalor, 2008).
Grassland stocking rate
N advice
(kg/ha N)
(LU/ha)*
(kg/ha)
≤ 90
<1.1
40
110
1.3
75
130
1.5
111
140
1.6
122
150
1.8
141
160
1.9
168
170
2.0
201
180
2.1
216
190
2.2
237
200
2.4
275
210
2.5
306
≥210
>2.5
279
*Based on annual nutrient excretion rates for livestock, as specified in SI 610 of 2010, e.g. a
dairy cow producing 85 kg N/year.
The N recommendations for cut grass are influenced by two main factors, the number of cuts taken
each year and the grazing history. They involve the following four stages:
i.
N application rates are advised for first and subsequent cuts of silage and for hay
(Table 11).
ii.
An adjustment is made for N that is applied at early grazing; “assume that 20% of this
remains available for first cut silage”
iii.
An adjustment is made for swards less than 4 years old; “an extra 25 kg/ha may be
used where necessary for establishment of a good ryegrass sward if pasture is less
than 4 years old”
iv.
An adjustments is made for SNS from grazing in the previous year; “where silage fields
were grazed rather than cut in the previous year, apply 100 kg/ha for first cut, and 85
kg/ha for second and subsequent cuts”.
45
Table 10. Suggested timing of available nitrogen (N) applications for swards grazed by bovines at
various stocking rates (Coulter and Lalor, 2008).
Stocking rate
(kg N/ha)
N rates (kg/ha) for approximate application dates
Jan/
Mar
Apr
May
Jun
Jul
Aug
Total N rate
Sep
(kg/ha)
Feb
≤ 90
25
40
110
15
30
15
15
75
130
28
35
25
25
111
140
28
35
25
25
17
122
150
29
44
26
26
17
141
160
29
44
35
35
26
168
170
34
53
42
42
31
201
180
32
32
48
38
38
28
216
190
31
41
54
37
37
37
237
200
30
53
53
37
37
37
27
275
210
31
54
54
56
37
37
37
306
≥210
32
55
55
38
38
38
28
279
At stocking rates above 210 kg/ha N, N advice is constrained by SI 610 of 2010.
Table 11. ‘Available nitrogen (N)’ rates for cut swards within the ROI grassland fertiliser
recommendations (Coulter and Lalor, 2008).
Crop
N application rate (kg/ha)
Silage: first cut
125
Silage: second or subsequent cut
100
Hay
65-80
For both cut and grazed grass recommendations, the contribution of crop available N from organic
manures is deducted from the advised total N rates to determine the amount of manufactured
fertiliser N to apply. A final check is then conducted to ensure that total N recommendations
(organic and manufactured N fertiliser) are compliant with maximum levels permitted within SI 610
of 2010.
5.3.
Fertiliser N recommendations in England and Wales – the “Fertiliser Manual
(RB209)”
The “Fertiliser Manual (RB209)” adopted a markedly different approach to grassland N fertiliser
recommendations compared with previous editions. Previous editions had used the economic
optimum to set N application rates based on the price of N fertiliser relative to the value of grass
46
dry matter with the break-even ratio (the amount of grass DM needed to pay for a kg of fertiliser N
applied) arbitrarily set at 7.5 or 10, depending on the livestock and grassland management system;
within the 7th edition N7.5 (the economic optimum N rate at a break-even ratio of 7.5) was adopted
for dairy grazing systems and cut grass, while N10 was adopted for beef and sheep grazing
systems. The new ‘systems’ approach adopted in the 8th edition (“Fertiliser Manual (RB209)”) was
based on the need to supply sufficient home-grown forage to meet the needs of a wide range of
livestock production systems at different levels of management intensity, defined in terms of
stocking rate, concentrate use and, for dairy systems, level of milk production. However, the
economic optimum concept was retained with Nopt set at N10 under cutting and grazing for all
livestock types and N10 values determined for Very good/Good, Average and Poor/Very poor GGC,
based on a selection of N response trials carried out in the period 1970 to 1993.
The N recommendations for established dairy grassland in England and Wales (Defra, 2010)
involved the following steps:
i.
Determine the GGC based on soil type (categorised by water holding capacity) and
Average Annual Summer Rainfall; to provide an indication of summer water supply.
ii. Use the GGC to select the relevant whole season total N requirement table.
iii. Select the target milk yield per cow per year.
iv. Calculate and select the dry matter tonnage of concentrated fed per cow per year.
v. Calculate and select the stocking rate in terms of livestock units (LU)/ha; the number of
livestock as LU (lactating cows plus followers on average on farm for the year) divided by
the total area of grass and forage crops, including forage maize.
vi. If the livestock system does not match any of the milk yield – concentrate use – stocking
rate categories within the Table, interpolate between stocking rate and concentrate values
by assuming a proportional difference in N requirement between values.
vii. Determine the SNS for the field; taking into account clover content and previous cropping,
grassland management (cut or grazed), N fertiliser use and organic manure applications.
viii. Adjust the whole season total N requirement according to the SNS (increase total fertiliser
N inputs by 30 kg N/ha in low SNS situations and decrease by 30 kg N/ha in high SNS
situations) and the clover content of the sward (estimated to supply 180-300 kg N/ha,
depending on clover content; 20-60%).
ix. Split the total N requirement for cut and grazed swards into 3-6 applications over the
growing season according to the recommended splits provided (e.g. for grass silage; 40%
for first cut (could be split further, Feb-Mar 15%, April 25%), 35% for second cut (could be
split further, May 20%, June 15%); 25% for subsequent cuts (could be split further, July
15%, August 10%).
47
x. An adjustment made for cutting after early spring grazing; reduce the 1st cut
recommendation by 25 kg N/ha.
The N recommendations for established grassland under beef production in England and Wales
(Defra, 2010) involved the following steps:
i.
Determine the GGC based on soil type (categorised by water holding capacity) and
Average Annual Summer Rainfall.
ii. Use the GGC to select the relevant whole season total N requirement table.
iii. Select the intensity of the system based on the housing period and amount of concentrate
fed through the winter housing season (e.g. the intensively and moderately grazed beef
systems are based on a typical housing period of 170 days, with concentrate use of 0.2 to
0.4 t/animal/yr).
iv. Calculate and select the dry matter tonnage of concentrated fed per cow per year.
v. Calculate and select the stocking rate in terms of livestock units (LU)/ha; the number of
livestock as LU (beef cattle plus young stock on average on farm for the year) divided by
the total area of grass and forage crops, including forage maize.
vi. If the livestock system does not match any of the intensity – concentrate use – stocking rate
categories within the Table, interpolate between stocking rate values by assuming a
proportional difference in N requirement between two values.
vii. Determine the SNS for the field; taking into account clover content and previous cropping,
grassland management (cut or grazed), N fertiliser use and organic manure applications.
viii. Adjust the whole season total nitrogen requirement according to the SNS and the clover
content of the sward.
ix. Split the total N requirement into 3-6 applications over the growing season according to the
recommended splits provided.
x. An adjustment made for cutting after early spring grazing; reduce the 1st cut
recommendation by 25 kg N/ha.
The N recommendations for established grassland under sheep production in England and Wales
(Defra, 2010) involved the following steps:
i.
Determine the GGC based on soil type (categorised by water holding capacity) and
Average Annual Summer Rainfall.
ii. Use the GGC to select the relevant whole season total N requirement table.
iii. Select the intensity of the system based on whether ewes are fed concentrates and
conserved forage during lambing (ewes fed 1.0 kg concentrate/ewe/d are in an intensively
grazed system and those fed 0.5 kg/ewe/d are moderately grazed).
48
iv. Calculate and select the stocking rate in terms of livestock units (LU)/ha; the number of
livestock as LU (ewes and lambs on average on farm for the year) divided by the total area
of grass and forage crops.
v. If the livestock system does not match any of the intensity – stocking rate categories within
the Table, interpolate between stocking rate values by assuming a proportional difference
in N requirement between two values.
vi. Determine the SNS for the field.
vii. Adjust the whole season total nitrogen requirement according to the SNS and the clover
content of the sward.
viii. Split the total N requirement into 3-6 applications over the growing season according to the
recommended splits provided.
ix. An adjustment made for cutting after early spring grazing; reduce the 1st cut
recommendation by 25 kg N/ha.
5.4.
Comparison of N recommendation systems
Dale et al. (2013) provided a useful summary of comparisons between the grassland N
recommendation systems in ROI, Scotland and England and Wales (
Table 12). They noted that, while there are clear differences in complexity between the systems,
there are also similarities in terms of the components used:

All three systems contain separate recommendations for grazed and cut swards, and
recommendations are adjusted to reflect different levels of production ‘intensity’. However,
only the SRUC Technical Note uses grassland management (defoliation sequences) as a
measure of production intensity and only the manual (RB209) takes account of milk yield (in
dairy systems) and concentrate inputs (in dairy and beef systems).

Both the “Fertiliser Manual (RB209)” and SRUC TN652 use Grass Growth Classes to
adjust N recommendations, with the manual using three GGC’s and TN652 adjusting N
rates on a five point scale. However, the manual and TN652 make opposite adjustments in
N rate to take account of GGC; the manual increases N rates for the poor growth class to
take account of lower N use efficiency in drier/cooler areas (e.g. to provide 7 t DM/ha in
intensively cut grass systems, N rates are increased by 140 kg N/ha for a Poor/Very poor
GGC site compared to a Good/Very good GGC site), while TN652 reduces N rates by 10
kg N/ha per site class resulting in N recommendations for site class 5 that are 40 kg N/ha
lower than site class 1. The ROI recommendations provide recommendations for soils with
an “average soil-N fertility”, stating that lower N rates may be used on high soil-N fertility
sites, and higher rates on low soil-N fertility sites where stocking rate is below 200 kg N/ha
(2.4 LU/ha); they therefore use a similar logic to the manual. It could be argued that Grass
Growth Classes are only justified where there are clear differences in N response between
49
classes and there are also regions of contrasting water supply (based on soil type and
summer rainfall) within a territory. Average annual rainfall (1981-2010) in ROI ranges from
1,000-1,400 mm in the west (with rainfall exceeding 2,000 mm in some mountainous areas)
to 750-1,000 mm in the east (Irish Meteorological Office). In Scotland the range within
productive grassland areas (i.e. lower altitudes) is from 2,000-3,000 mm in the west to 7001,000 mm in the east; and in England and Wales from 1,000-2,000 mm in the west to 600800 mm in the east (Met Office). The need for Grass Growth Classes to represent regions
with contrasting grass growth potential may therefore be greatest in England and Wales.

TN652 does not take account of SNS, while the manual specifies three Soil Nitrogen
Supply categories based on current clover content and cropping, grassland management
(cut or grazed), N fertiliser use and organic manure applications in previous years. The ROI
recommendations highlight that where cut swards were grazed in the previous year, the N
recommendations for cutting should be reduced by 25 kg N/ha for first cut and 15 kg N/ha
for second and subsequent cuts.

Only the manual takes into account substitution rate, forage utilisation (spoilage) and grass
and silage energy content. The manual also includes a wide range of systems, with
stocking rates ranging from 1.5 to 4.0 LU/ha, compared with <1.1 to >2.5 LU/ha for the
grazing recommendations in the ROI. However, the manual (RB209) provides
recommendations for high levels of purchased feeds (e.g. 4.4 t concentrate/cow/year at a
stocking rates of 3.0-4.0 LU/ha), which exposes dairy producers to market volatility in terms
of feed and milk prices (e.g. Mihailescu et al., 2015b).
Validation of N response curves covered in section 4.1
There are a number of other factors that influence grass growth and its response to fertiliser N that
need to be considered as part of an integrated approach to soil and nutrient management. These
include the influence of defoliation processes, treading and compaction, as well as the effects of
faecal and urine returns. However, while some of these factors, such as nutrient cycling could be
incorporated into the energy or nutrient model that underpin the recommendations, most of them
are not key input factors to be included in a N recommendation system for grassland advisers and
farmers.
50
Table 12. A comparison of the components included in the nitrogen (N) fertiliser recommendations
currently adopted in Scotland the Republic of Ireland and within England and Wales (Dale et al.,
2013).
Scotland
Republic of Ireland
England and Wales
(RB209 8th Edition)
Grass Growth/Site class



Soil Nitrogen Supply

1

Grass management



Stocking rate



Milk yield



Concentrate input



Separate recommendations for
cut and grazed swards



Expected grass requirements
calculated and expected herbage
yields indicated



Substitution rate, forage
utilisation, grass and silage
energy content included in
calculation of forage requirements



Intensity of production system –
1
for cut swards only, although “lower N rates” are advised for grazed swards with “above average natural
fertility”.
Use of the ‘energy model’ – influence of principal factors on N recommendations
There have been a number of criticisms of the N recommendations in the “Fertiliser Manual
(RB209)” both in terms of the complexity and number of adjustments required to generate a N
recommendation for each cut or grazing, and in the energy model or ‘back calculation’ method that
underpins the recommendations (Dale et al, 2013; Defra project IF01121). Within the energy
model, the fundamental DM yield response of grass to N fertiliser is based on field trials carried out
from 1970 and 1993 and the same response curves are used for cut and grazed grass (see
Section 4.1). The amount of N fertiliser needed to meet livestock energy requirements, therefore,
does not take account of the improved production performance of modern perennial ryegrass
varieties or the contrasting potential for DM offtake of infrequently cut and frequently grazed
swards. The following assumptions used within the “Fertiliser Manual (RB209)” are highlighted by
Dale et al. (2013); each factor has implications for N fertiliser requirements, although some factors
in combination would cancel each other out:
51

Improved production performance of modern perennial ryegrass varieties – would reduce N
fertiliser recommendations

Lower DM yield potential of grazed swards compared with cut swards – would increase N
fertiliser recommendations for grazed swards

Underestimation of ME required for maintenance - would increase N fertiliser
recommendations

Overestimation of the average liveweight of dairy cows - would reduce N fertiliser
recommendations

Underestimation of the energy required to produce a litre of milk - would increase N
fertiliser recommendations

No account of energy associated with live weight change during lactation, growth, or
pregnancy - would increase N fertiliser recommendations

Underestimation of the ME content of grass silage – would increase N fertiliser
recommendations

Overestimation of the utilisation of forage at grazing (80%, cf. 50-70%; e.g. Mayne et al.,
2002; Ganche et al., 2012) - would increase N fertiliser recommendations
While each of the above factors in isolation can have a significant effect on the N fertiliser levels
recommended, when combined they have the effect of cancelling each other out and the overall
balance would depend on the specific values selected for each factor. So, while the energy model
could be criticised as being overly simplistic it could also be argued that a more complex approach
that includes a greater number of fine adjustments could make revision of the model more
demanding with a relatively small improvement in the accuracy of the recommendations provided.
Dale et al. (2013) also used CAFRE Benchmarking data (from 2009-2012) and data from 12 dairy
farms (monitored in the Vision II and DAIRYMAN projects) to compare the RB209 8th edition N
recommendations with actual milk yield – concentrate input – stocking rate scenarios and current N
usage on silage fields in N.I. They found that the “Fertiliser Manual (RB209)”:

Overestimates cut grass N requirements (by 95 kg N/ha on average) for lower intensity milk
production systems; and both cut and grazed N recommendations in scenarios where
stocking rates are high and concentrate feed levels are low

Underestimates cut grass N requirements (by 80 kg N/ha on average) for higher intensity
production systems; and both cut and grazed N recommendations when stocking rates are
<2.0 cows/ha and concentrate levels are at the upper end of the range
52
These findings have important implications for the validity of the 8th edition recommendation
system under certain scenarios. However, to revise the systems approach, incorporating a more
‘accurate’ energy model would require a significant amount of research including sensitivity
analysis to assess the consequences in terms of the resulting N recommendations. It may be more
pragmatic to adopt a simplified recommendation system that avoids the potential pitfalls of a
detailed systems approach.
Overview of different recommendation systems and discussion of ‘systems’ approach to nutrient
management.
A ‘systems’ approach encourages farmers and advisers to consider the amount of purchased feed
and manufactured fertiliser that is used to achieve a desired level of liveweight gain or milk yield.
Clearly, the amount of concentrate and manufactured fertiliser purchased has implications for the
nutrient use efficiency and net profitability of the system, as well as the resilience to market
pressures such as relative changes in the price of meat, milk, feed and fertiliser. For example,
Mihailescu et al. (2015b) carried out a 3-year study in Ireland, assessing the impact of N and P use
efficiency on 19 intensive grass-based dairy farms. The study found that mean net profit increased
with mean milk receipts and decreased with mean expenditure on manufactured fertilisers,
implying that increasing milk receipts while optimising the use of manufactured fertiliser can be an
effective strategy to increase net profit. The study also concluded that higher input systems were
more vulnerable in periods of low milk prices and emphasised the importance of optimising inputs
to improve economic sustainability. Eight farms exceeding the limit of 2 livestock units (LU)/ha,
imposed through the Nitrates Directive, had 1.6 times higher net profit compared with the
remainder. Nevertheless, the results indicated that Irish dairy farms, as low-input production
systems, have the potential to improve both economic production (as indicated by net profit per
hectare) and environmental sustainability (as indicated by N and P balances per hectare, N and P
use efficiencies and litres of milk produced per kilogram of imported N). The importance of the
balance between grass forage production and concentrate use is also supported by Ryan et al.
(2011) who evaluated nitrogen efficiency as a key indicator of economically sustainable production
on grass-based and high-concentrate dairy production systems in Ireland. They reported that as N
concentrate increased, N surplus per hectare increased and N use efficiency per hectare
decreased.
Mihailescu et al. (2014) reported improvements in nutrient management within grassland systems
in the Republic of Ireland. Case studies of intensive grass-based dairy farms, before and after the
implementation of good agricultural practice (GAP) regulations in 2006, indicated that N surplus
both per ha and per kg of milk solids had decreased by 40% and 32%, respectively (Mihailescu et
al. 2014). The reductions in N-surplus were achieved through a reduction in manufactured N
fertiliser inputs and increased spring application of organic manures. Following the implementation
53
of GAP regulation, mean N-surplus was 175 kg N/ha, which was lower than the mean for dairy
farms in northern and continental Europe of 224 kg N/ha. The mean NUE of 0.23 was similar to the
European mean, reflecting the low input/ low output systems typically adopted in ROI, involving
seasonal milk production, low use of concentrates, imported feed and forages, high use of grazed
grass and lower milk yields per hectare (Mihailescu et al. 2014).
Thus, these studies highlight the importance for N fertiliser recommendations to provide clear
guidance on the amount of DM yield that can be produced at different levels of N input. Ideally,
recommendations should direct users towards profitable systems that are resilient within a variety
of market conditions.
5.5.
Options for presenting nitrogen recommendations for grassland
Current and previous recommendation systems in the UK and ROI provide a number of
approaches to presenting N recommendations for grassland. The options for presenting grassland
N recommendations in England and Wales include:

Retain the systems approach, including the current energy model that underpins the current
recommendations, and incorporate a table that provides indicative DM yields according to
N fertiliser rate for each GGC to introduce the concept of growing enough grass to meet the
energy needs of any given system (Appendix II). The recommendations could also include
a flow chart to take users through the process of generating a N recommendation; and
worked examples and case studies to illustrate how the process works for low, medium and
high output systems. This approach provides recommendations for a wide range of
intensities and systems, and includes factors such as forage utilisation and forage energy
content to calculate grass forage and therefore N requirement. The systems approach
provides flexibility in matching recommendations to the production system on each farm.
However, the resultant recommendations are much more complicated than other systems.
Given that many producers are not able to fit their system into the wide range of milk yield,
concentrate use and stocking rate combinations provided in the 8th edition (Defra project
IF01121), a narrower range of viable systems could be presented to illustrate a narrower
range of possible combinations with fewer rows per Table. This would avoid the implication
that the range of systems presented are exhaustive.

A Table that presents indicative DM yield to represent the range of productivity in swards
with differing proportions of modern perennial ryegrass varieties under a cutting regime at
various levels of N fertiliser use and for three Grass Growth Classes.

N recommendations under a grazing system by stocking rate or indicative DM yield,
assuming low concentrate use; similar to the grazing recommendations used in ROI
54
(Coulter and Lalor, 2008). N rates can be adjusted from mid-season onwards according to
grass growth, summer rainfall and livestock requirements. Such an approach is simple and
easy to interpret, and is appropriate for production systems based on spring calving and
grazed grass.

N recommendations for a number of selected defoliation or management sequences (i.e.
similar to the approach for Scotland) for specific levels of production, based on indicative
target DM yields. This approach allows users to select a level of N use to achieve the DM
yields required for their system, is easy to interpret, clearly sets out how much N to apply
for each cut or grazing cycle, and offers sufficient detail to allow recommendations to be
tailored to different intensities of production.

Specific grass silage crop recommendations presented by cut for a given GGC and SNS
with adjustments made according to rainfall from mid-season onwards. Different
recommendations could be provided for different levels of production (Appendix II).
Recommendation tables should be complemented by text that outlines the principles of nutrient
management for grassland and clearly specifies the key factors and general issues to consider. A
number of options for presenting grassland N recommendations are provided in Appendix II.
6.
Response of forage crops to applied nutrients
6.1.
Response to nitrogen
Maize
The “Fertiliser Manual (RB209)” recommends 150 kg N/ha of fertiliser N at a SNS of zero, i.e.
around 170-210 kg N/ha in terms of N supply. An inadequate supply of N results in less starch for
storage in the grain, and reduced yield and protein (PDA, 2008). However, if the nitrogen supply is
too high, leaf growth is excessive, increasing the proportion of leaf and stem to grain, reducing the
starch content of silage, delaying maturity and potentially resulting in lodged crops (PDA, 2008;
Coulter and Lalor, 2008). Applying too much N can therefore be more damaging than applying too
little (Coulter and Lalor, 2008).
At 100 sites over five years (2010-2014), Jewkes et al. (2013) measured soil mineral N (SMN) and
mineralisable N in the spring prior to manure applications; and dry matter yield and N offtake
(stems and cobs) at harvest to determine the range of SNS, N offtake, DM yield and overall N
balance (SMN and mineralisable N plus added N fertiliser, minus total plant N) values in forage
maize crops. The relationship between N balance and total plant N was linear, with the amount of
N in the crop at harvest increasing with increasing negative N balance. At a zero N balance, the
crop offtake was 189 kg N/ha (+/- 40 kg/ha), which corresponded to a DM yield of 15.7 t DM/ha (+/55
1.8 t DM/ha). This was below the level at which cob production (in which starch/energy production
is concentrated) is compromised by stover/leaf growth. Plotting a simplified N balance (i.e. SMN
plus mineralisable N minus plant N) against DM yield indicated a mean N requirement of 95 kg
N/ha (as manufactured N fertiliser and/or manure crop available N) for a DM yield of 16 t/ha at zero
balance.
Corcoran et al. (2016) investigated the effects of degradable plastic mulch (green and yellow with
degradation scores of 3 and 5 respectively), N fertiliser rate (0, 50, 100 and 150 kg N/ha) and N
application timing (100% at sowing or 50% at sowing and 50% at growth stage V6-V8) on forage
maize yield and composition. The green plastic mulch treatment resulted in greater whole-crop
yield (P<0.05) (13.5 cf. 12.7 t DM/ha), and crop nitrogen uptake (P<0.05) compared to the yellow
mulch, although the yellow mulch had a higher starch content (P<0.05; 372 cf. 327 g/kg DM). The
split application of N 50% at sowing and 50% at V6-V8 increased starch concentration (P<0.001),
but decreased nitrogen use efficiency and whole-crop yield (P<0.05). The split application of N
conferred no overall benefit compared to 100% application of N at sowing. The highest grain yields
were achieved at 100 kg N/ha under yellow mulch (7.0 t DM/ha) and at 150 kg N/ha under green
mulch (7.7 t DM/ha). The highest whole-crop yields were achieved at 150 kg N/ha for both yellow
(13.4 t DM/ha) and green (15.1 t DM/ha) plastic mulch.
Kayser et al. (2011) reported findings from a 4-year experiment on the response of forage maize to
N application at N-rates (0, 80, 160 and 240 kg N/ha/yr), applied either as manufactured fertiliser,
cattle slurry or pig slurry. Over the 4-years, yields and N-offtakes at 0N were high (mean of c.13.5
t/ha and c.150 kg N/ha, respectively), indicating high SNS. Responses to N-input were small with
apparent N recoveries of 14-22% for manures and mineral fertilisers. The study concluded, that
care should be taken when growing maize on soils with high potential N mineralisation.
Lynch et al. (2013) assessed the effect of N application rate (c.30 and c.170 kg N/ha, supplied as
30 kg N/ha as crop available N from cattle slurry and the remainder as calcium ammonium nitrate),
harvest date and cultivar on maize yield and the quality of whole-crop, cob and stover silages.
Overall, the study found no effect of N application rate on DM yield, nutritive value or ensiling
characteristics of maize whole-crop or cob silage. However, the higher N-rate (c.170 kg N/ha)
significantly increased stover DM yield (P<0.05) compared to the lower N-rate (c.30 kg N/ha).
Whole-crop and stover harvested later had a lower digestible content and underwent a more
restricted fermentation compared to silages produced from maize harvested earlier.
As part of Defra project WQ0140 (Competitive maize cultivation project with reduced
environmental impact) the response of maize to N application rate (0, 40, 80, 120, 160 and 200 kg
N/ha) was measured at two sites one in Nottinghamshire the other in Norfolk in 2014, on sandy
56
loam textured soils. At Nottinghamshire, yield response to N application rate was low, giving an Nopt of c.20 kg N/ha and DM yield of 17.5 t/ha, at a BER of 6.5. While at the Norfolk site, maize
yields increased linearly with N application rate and N-opt was not reached; the highest N rate (200
kg N/ha) resulted in a mean DM yield of 18.8 t/ha. The high yields achieved at both experimental
sites were typical of the year, due to the ideal maize growing conditions in 2014.
The above results indicate that no change is needed to the N recommendations for forage maize in
the 8th edition of RB209.
Forage rape
Keogh et al. (2012) assessed the impact of N application on forage rape yield in a field study at
three sites at Moorepark in Ireland. Overall, it was found that forage rape yield increased up to the
maximum N rate applied (120 kg N/ha), when sown in early August, giving a mean yield (across the
3 experiments) of 4.9 t/ha. The study concluded that the optimal sowing time for forage rape in
Ireland was early August. The results for forage rape indicate that there may be a case for a small
increase in N recommendations in zero SNS Index situations, but there is insufficient data from
England and Wales to justify this.
Stubble turnips
Keogh et al. (2012) also assessed the impact of N application on stubble turnip yields at two sites
of contrasting fertility (SNS). Stubble turnips sown in early to mid-August, showed less of a
response beyond 40 kg N/ha giving a mean yield (of the two sowing dates) of 4.1 t/ha at the fertile
site. However at the less fertile site, yield increased up to the maximum N-rate applied (120 kg
N/ha), giving a mean yield (at the two earlier sowing dates) of 4.7 t/ha. The results for stubble
turnips indicate that there may be a case for a small increase in N recommendations in zero SNS
Index situations, but there is insufficient data from England and Wales to justify this.
Forage triticale
Knapowski et al. (2012) assessed the effect of N application rate (80 or 120 kg N/ha) and foliar
application of zinc (Zn) (applied at 0, 0.1 and 0.3 kg Zn/ha) on grain Zn and copper (Cu) content of
triticale. Overall, it was found that an N application of 120 kg N/ha resulted in a significant increase
in the Zn content and a decrease in the Cu concentration in grain compared to applications of 80
kg N/ha. Furthermore, applying foliar Zn at all rates resulted in significant increases in Zn content
and a decrease in the copper concentration in triticale grain. Wojtkowiak et al. (2014) found that
higher (120 kg N/ha) N applications to triticale, as urea in 2 splits, (at tilling and a foliar application
at stem extension) resulted in higher concentrations of P, calcium (Ca) and magnesium (Mg). The
accumulation of glutenins typically increased in response to higher (120 kg N/ha) doses of N.
57
Innovate UK project 101093, which was reported as part of AHDB Cereals & Oilseeds project
3699, investigated the N response of modern triticale crops. Clarke et al. (2016) found no
significant difference in the N requirements of winter wheat and triticale (see WP4 report).
However, it was recognised that triticale has a greater lodging risk than wheat, so less N may be
required in situations of high lodging risk. Nevertheless, the results indicate that RB209 (8th edition)
N recommendations for triticale (which are about 100 kg/ha lower for triticale than for wheat) were
insufficient to meet crop demand. It is therefore proposed that winter triticale grown for grain should
use the same N recommendation table as winter wheat. However, forage triticale is harvested
earlier than grain triticale. Based on harvest at early milky development (GS71; mid/end June) and
the fact that total N uptake by GS61 was not significantly different between the triticale and wheat
varieties (Clarke et al., 2016), it is proposed that the new triticale N recommendations (i.e. winter
wheat recommendations) should be reduce by 50 kg N/ha (assuming 30 kg N/ha less uptake and
60% fertiliser efficiency). On medium soils at SNS Index 1, this would equate to a N
recommendation of 170 kg N/ha.
Whole-crop wheat
No new data was available to review the N requirements of whole-crop wheat. However, it was
agreed at the Livestock TWG on 26th April 2016 that N recommendations for whole-crop wheat
should be the same as wheat grown for grain with no adjustments made to account for harvest
date, as whole-crop wheat is generally cut later than forage triticale. For example, fermented
whole-crop is cut at soft dough stage (GS85), while high dry matter whole-crop is cut closer to fully
ripe (GS87-89).
Kale
Understanding how nitrate and nitrite accumulate in forage crops is critical, in order to prevent
feeding livestock forage containing high nitrate levels which can negatively impact on animal
health, i.e. through nitrite poisoning. The timing of harvest can have a significant effect on the Ncontent of kale; Korus and Lisiewska (2009) reported that later cuts of kale contained between
17% more total-N and 8% more protein-N than earlier cuts, with mean total-N and protein-N (over
2 experimental years) of 0.58 and 0.52 g /100g, respectively. Furthermore, nitrate and nitrite
concentration were reduced by 83% (294 mg /kg) and 46% (0.22 mg/kg) in later cuts. While no
yield data was available from this study, the findings can be used to provide advice on the time
interval between N fertiliser applications and harvesting or grazing of kale for forage.
Swede and kale grazed in situ
No new data was available on nutrient requirements for crops grazed in situ, so it was proposed
that the N recommendations for grazed crops should be the same as the recommendations for
harvested crops. However, using grazing return principles, the SNS following swede and kale
58
grazed in situ should be increased by one level relative to harvested crops, as is the case in The
Green Book (Coulter and Lalor, 2008). This was supported by the Livestock TWG on 26th April
2016.
6.2.
Response to sulphur
Maize
Trials data indicate that some forage maize crops respond to sulphur applications. In a replicated
field plot trial conducted at Henfaes Research Centre, near Bangor, a sulphur application of 18 kg
SO3/ha yielded 1.9 t DM/ha more than where sulphur was not applied (13.8 t/ha cf. 11.9 t/ha;
P<0.05). However, the higher rates of S application (36 and 54 kg SO3/ha) had no effect. Nitrogen
was applied at 109 kg N/ha, and phosphate and potash according to RB209 recommendations.
At a Maize Growers association (MGA) site in Oxfordshire (sandy silt loam; pH 6.2, P Index 4, K
Index 2; previous crop: tomatoes) a trial was carried out to investigate the yield and quality
response of forage maize to sulphur at three levels of N (17.5, 35 and 50 kg N/ha) and three
corresponding levels of S (50, 100 and 150 kg SO3/ha); using ammonium nitrate to provide the N
treatments and ammonium sulphate to provide the N + S treatments. The addition of S increased
DM yields at each rate of N (17.5 kg N/ha, with/without 50 kg SO3/ha; 35 kg N/ha with/without 100
kg SO3/ha; and 53 kg N/ha with/without 150 kg SO3/ha). DM yield increases of 33% were
measured at 18 kg N/ha from 50 kg SO3/ha; 15% at 35 kg N/ha from 100 kg SO3/ha; and 15% at
53 kg N/ha from 150 kg SO3/ha (P <0.05). The highest N and S rates also increased starch and
protein content (P <0.05). However, at a similar trial in Cheshire there was no DM yield response
to applied N or S fertiliser.
Wortmann et al. (2009) estimated the forage maize physiological requirement for S was 60 kg
SO3/ha for a crop of 15 t/ha. This is the amount that needs to be supplied from all sources; air, soil
and applied fertiliser. The authors reported no DM yield response to applied S fertiliser and
attributed this to adequate SOM concentrations and associated mineralisation of S at the sites
investigated. Webb et al. (2015) reported that sulphur deposition in England and Wales, net of
leaching was 8.5-12 kg SO3/ha in 2009-11 and was estimated to be 4.0-7.5 kg SO3/ha by 2020.
The “Fertiliser Manual (RB209)” and The Green Book (Coulter and Lalor, 2008) do not recommend
any sulphur on forage maize as a low protein crop grown primarily for starch production. Indeed,
many forage maize crops receive livestock manures or other organic materials (e.g. digestate) that
contain both sulphur and nitrogen and it could be considered that the modest requirement of forage
maize for S could be met by this source. Nevertheless, a number of trials have measured a forage
maize DM yield response to the application of S fertiliser. Therefore, on land that does not receive
regular applications of organic manure, and with declining rates of S deposition, it would be
59
prudent to follow the guidelines proposed by Cussans et al. (2007) for wheat, and apply 20-50 kg
SO3/ha to forage maize crops (applied in seedbed or at least before two leaf stage) grown on
sandy soils; loamy and coarse silty soils in areas with > 175 mm overwinter rainfall; and on clay,
fine silty or peat soils in areas with >375 mm overwinter rainfall.
Forage rye and triticale
No new data was available on the response of forage rye and triticale to sulphur applications. In
the absence of any research to quantify the S response from oats, rye and triticale, it is
recommended that the 25-50 kg/ha SO3 recommendation for winter wheat is applied to all winter
and spring sown cereals crops. However, not all cereal crops will respond to S. The risk matrix
developed by Cussans et al. (2007) and published in AHDB Cereals & Oilseeds Information Sheet
28 should therefore be included in a revised RB209.
Forage rape
No new data was available on the response of forage rape to sulphur applications. However, the
recommendations for oilseed rape (OSR) can be used as a comparison. Recent work on winter
OSR supports the current RB209 recommended rate of 50-75 kg/ha SO3 (see WP4 report). Given
the unavailability of data on forage rape it is recommend that the 50-75 kg/ha SO3
recommendation for winter OSR is applied to forage rape.
Other forage crops
In view of the unavailability of data on the response of other forage crops to sulphur it is
recommended that the risk matrix developed by Cussans et al. (2007) is adopted with a
recommended application of 25-50 kg/ha SO3 in higher S deficiency risk situations and where
organic manures are not applied prior to the growing season; or organic manures have not been
regularly applied in previous years.
6.3.
Response to phosphate and potash
The K2O and P2O5 requirements of different forage crops (whole-crop wheat, Italian ryegrass,
perennial ryegrass/white clover, fodder beet, kale and maize) were assessed in a 3-year field trial
(PDA leaflet 26, 2007) with nutrients supplied by applying a combination of manure and
manufactured fertiliser. Overall, the results demonstrated that more K2O was removed than N,
while K2O removal was 3-6 times greater than P2O5 removal (Table 13). Overall, P2O5 offtakes
were slightly lower and K2O offtakes higher than standard values (Table 13). For fodder beet, there
was a large difference between measured K2O offtake (4.0 kg/t FW) and standard figures (1.7 kg/t
FW for sugar beet; roots only), indicating that further evidence is required to refine standard values
for K2O removal by fodder beet.
60
Table 13. Amount of phosphate and potash removed by forage crops (taken from PDA leaflet 26,
2007b).
Measured values
Standard values
Phosphate
Potash
Phosphate
Potash
Whole-crop wheat
1.8
5.4
n/d
n/d
Italian ryegrass
1.4
6.3
1.7
6.0
Perennial ryegrass/white clover
1.3
6.8
1.7
6.0
Fodder beet
0.7
4.0
0.8
1.7
Kale
0.9
5.0
1.2
5.0
Maize (silage)
1.7
6.5
1.4
4.4
Overall, due to the limited availability of new data to review the phosphate and potash
requirements of forage crops, it is proposed that no changes be made to the recommendations
apart from introducing fodder beet (roots only) offtake values.
Swede and kale grazed in situ
Using grazing return principles, it was proposed that the P2O5 and K2O requirement on grazed
crops should be reduced by 30-60 kg/ha relative to harvested crops, due to less nutrient offtake.
This was supported by the Livestock TWG on 26th April 2016.
6.4.
Conclusions for forage crop fertiliser recommendations
Nitrogen

No change is needed to the N recommendations for forage maize in the 8th edition of
RB209.

There may be a case for a small increase in N recommendations for forage rape and
stubble turnips in zero SNS Index situations, but there was insufficient data from England
and Wales to justify this.

It is proposed that the N recommendations for forage triticale should be 50 kg N/ha lower
than the new winter wheat recommendations to account for the earlier harvest date for
forage triticale, and assuming 30 kg N/ha less uptake and 60% fertiliser efficiency.

The N recommendations for whole-crop wheat should be the same as for wheat grown for
grain.

The SNS following swede and kale grazed in situ should be increased by one level relative
to harvested crops.
Sulphur
61

For non-Brassica forage crops it is recommended that on land not receiving regular
applications of organic manure or where sulphur-containing organic materials have not
been applied prior to the growing season, it would be prudent to follow the guidelines
proposed by Cussans et al. (2007) for wheat and apply 20-50 kg SO3/ha in higher risk
situations according to the soil type/overwinter rainfall guidelines.

For forage rape and kale grown on mineral soils, where organic manures have not been
applied regularly in previous years, it is recommended that 50-75 kg/ha SO3 be applied as
a sulphate containing fertiliser in late February to early March.
Other nutrients

Given the lack of available data on the response of forage crops to other nutrients, it is
proposed that no changes be made to the recommendations.

The one exception is that on swede and kale grazed in situ, phosphate and potash
recommendations should be reduced by 30-60 kg/ha relative to harvested crops, to reflect
lower nutrient offtake (i.e. grazing returns).
7.
Gaps in knowledge and future research required
Knowledge gaps are greatest for the fertiliser requirements of grazed grass due to limited
information on recycling of nutrients, spoilage and wastage at grazing with modern grass varieties;
the contribution of modern clover varieties to grass dry matter yields and SNS and how this varies
with weather conditions from year to year; contrasting grass DM yield response to N fertiliser of
older and more modern grass varieties and at lower and higher altitude sites; and the response of
forage crops to nutrients in a range of conditions (Table 14). A number of research papers and
projects support the knowledge gaps identified in Table 14. For example:
Grazed systems
There is limited evidence on the impact of grazing on nutrient recycling. Eriksen et al. (2015)
demonstrated that nitrate leaching losses can be high from grass-clover swards that are both
grazed and receiving N-inputs from manufactured fertiliser. Further work is required to link stocking
rate, grazing behaviour to nutrient recycling and herbage yields and quality.
Grass-clover swards
Peyraud et al. (2009) and Phelan et al. (2014) outlined a number of limitations surrounding the use
of legumes which require further research. It is difficult to predict DM production in grass-clover
swards and maintaining optimum clover contents in swards is challenging. In addition, legumes
can be difficult to conserve as silage or hay, and can increase the risk of bloat in livestock.
62
Phelan et al. (2015) outlined that further research into forage legumes, should take a “back to
basics” approach and focus on the effects of management practices including sowing dates,
regrowth/rest periods and post-grazing heights.
Impacts on SNS with a rotation
The importance of accounting for high mineralisation potential in crop rotations has been
highlighted in studies which have shown low N-fertiliser responses and high NO3-N leaching losses
from maize following grass (e.g. Eriksen et al., 2015 and Kayser et al., 2011). Further research is
required to understand the yield response of maize at varying SNS Indices and climatic conditions
in order to maximise nutrient-use efficiencies. Nevertheless, over-sowing maize with ryegrass can
be effective at reducing surplus soil mineral nitrogen content before the onset of over-winter
drainage, thereby reducing NO3-N leaching losses (e.g. Eriksen et al., 2015 and Defra project
WQ0140). Further research into over-sown maize is required, in order to provide clear guidance on
establishment methods, cover crop species and destruction techniques that do not have a negative
impact on maize yields and quality.
Forage crops
Korus and Lisiewska (2009) highlighted that further research is required to improve our
understanding of the accumulation of nitrate and nitrite in kale. Nitrate is as a source of nitrite,
which is formed in the rumen after ingestion. Nitrite poisoning is uncommon in store lambs, but
when the problem occurs, lamb losses can be high.
Nutrient-use efficiency
Mihailescu et al. (2014 & 2015a) reported that N and P use efficiency has increased on Irish grassbased dairy farms since the introduction of GAP regulations. Improvements have mainly arisen due
to a reduction in N and P fertiliser inputs and a shift towards spring application of organic manures,
thus indicating an increased awareness of the fertiliser value of organic manures and the
importance of accounting for nutrient inputs from manure when planning fertiliser applications. The
dominance of N-fertiliser inputs on Irish low input/ low output dairy farms means that improvement
to the NUE of fertiliser N and organic manures will be of central importance for reducing Nsurpluses and increasing NUE and productivity. Mihailescu et al. (2014) suggest improvements to
N-balances and NUE can be achieved through optimising management practices such as: nutrient
management planning, grazing management and grass utilisation and use of grass-clover swards.
Whilst, for P further improvements in PUE can be achieved by optimising P fertiliser application
and feed imports and improved on-farm P recycling.
63
Table 14. Summary of knowledge gaps relating to grass and forage crops, and suggestions for work
to address these gaps.
Future work
Level of
priority
Limited information on
recycling of nutrients, spoilage
and wastage at grazing
Grazing
experiments on
response to N
High
Cut grass N
Information on N response
across a range of conditions.
N response of old
and new grass
swards tested in a
range of GGC
and agro-climatic
conditions
High
Soil compaction
Limited information on the
effect of soil structural
degradation on grass and
grass/clover response to
nutrients
N response
experiments in
field situations on
soils in moderate
and poor
condition
High
Cut grass S
Information on S response,
particularly at first cut.
S response of old
and new grass
swards tested in a
range of GGC
and agro-climatic
conditions
High
Cut grass K
Information on potash offtake
values in cut grass
Collation of data
from various
sources including
future N response
experiments
High
Grass P
Information on grass response
at P Index 2 and case for
splitting into two sub-bands
Analysis of grass
herbage in
relation to soil
Olsen P and P
fertiliser
applications
Medium
Clover DM yield
and N fixing
Contribution of modern clover
varieties to grass dry matter
yields and SNS
Response of
modern grass and
clover varieties to
differential rates
of N
High
Clover DM yield
and older/younger
perennial
ryegrass swards
Production potential of grassclover swards
Comparison
between clover
swards with older
and modern grass
varieties
High
Clover
persistence in
grass-clover
swards
Techniques to increase the
persistence of clover in grassclover swards
Trials
investigating
contrasting
management
regimes
Medium
Forage crop
nutrients
Response of forage crops to
nutrients; and P and K offtake
values of modern varieties
N and S response
experiments
including
measurement of
nutrient offtakes
High
Area
Knowledge gaps
Grazed grass N
Relevant
work
underway
64
Future work
Level of
priority
Level of beneficial and toxic
substances in forage crops
N and S response
experiments to
integrate herbage
analysis
Medium
Yield response of
maize in varying
SNS and climatic
conditions
Effect of site history on forage
maize N and S requirements
and nutrient-use efficiencies
N and S response
experiments at a
variety of sites
High
P-fixing capacity
The range of P-fixing capacity
within grassland soils in
England and Wales
Literature search
to determine key
P-fixing properties
and soil mapping
High
Biostimulants
Effect on grass yield and
quality in UK conditions
Yield and quality
response
experiments
Medium
Area
Knowledge gaps
Accumulation of
nutrients in forage
rape and kale
8.
Relevant
work
underway
Conclusions
Nitrogen response of grass and clover
A limited number of N response experiments indicate that modern grass swards (1 to 10 years old)
are capable of producing higher grass DM yields than older grass swards. This has implications for
N grassland recommendations in terms of the amount of N to apply to produce a given level of
grass DM. Indicative DM yields provided in the N recommendations should reflect the N response
of older and younger grass varieties and the varying composition of grass swards. However, more
data is needed on the response of modern grass and grass-clover swards to nitrogen.
Sulphur recommendations for grass
The current grass sulphur recommendations appear sound, but it is recommended that the advice
to apply S fertiliser in the situations suggested by Cussans et al. (2007) is included. Given the
findings of Bailey (2016), it may also be sensible to remove the sentence “Deficiency at first cut is
less common but can occur on light sand and shallow soils”, while being wary that over supply of S
can be detrimental to animal health.
Phosphate and potash recommendations for grass
Work is needed in England and Wales to assess the P-fixing capacity of different soils to produce a
map of soil parent material types (as defined by soil associations) and the contrasting P response
of different crops. Research in ROI and NI indicates that Olsen P Index 2 could be split into a lower
and upper sub-band with grass P recommendations increased for the lower sub-band. Similar
research is required in England and Wales to determine the herbage P sufficiency status at
different levels of Olsen P and P fertiliser use. More data is needed on potash offtake in grass for
65
which the grassland management (primarily cutting date, K Index and nutrient applications) is
known.
Lime and soil sampling and analysis recommendations
There is no need to change the fundamental principles currently provided. However, the lime factor
will be incorporated into lime recommendation tables; and advice on GPS sampling for soil pH and
nutrient reserves, and the variable rate application of lime will be incorporated into the “Nutrient
Management Guide (RB209)”. Consideration should also be given to the inclusion of advice on the
use of seashell sand and its impact on soil pH.
Grazing and grassland management recommendations
There is currently not sufficient data on grass DM yields under grazing management in England
and Wales. The current approach that uses the same N response curves for cutting and grazing
should therefore be retained, but account will be taken of nutrient returns at grazing.
Grazing height can influence sward recovery, DM yield and overall productivity of grazing livestock
systems. However, there are a number of approaches within the industry and advice on herbage
mass at the start and end of each grazing cycle should not form part of nutrient management
recommendations.
Nutrient management guidelines should include advice on grassland and soil management related
to the risks of causing compaction or poaching. Readers will be signposted to industry advice on soil
structure in the “Principles of nutrient management” section.
Grass and silage quality recommendations
For first cut silage, manufactured N fertiliser application rates above 120 kg N/ha should be avoided,
where possible.
The importance of micronutrients for grass/clover productivity and animal health needs to be made
clear in the recommendations and options for correcting potential deficiencies provided, including
the use of micronutrients in fertilisers (e.g. selenium).
Research should be carried out to determine whether biostimulants can increase yield and quality of
grass/clover swards in England and Wales.
66
Forage crop N fertiliser recommendations
No change is needed to the N recommendations for forage maize in the 8th edition of RB209. There
may be a case for a small increase in N recommendations for forage rape and stubble turnips in zero
SNS Index situations, but there was insufficient data from England and Wales to justify this.
It is proposed that the N recommendations for forage triticale should be 50 kg N/ha lower than the
new winter wheat recommendations to account for the earlier harvest date for forage triticale, and
assuming 30 kg N/ha less uptake and 60% fertiliser efficiency. Whole-crop wheat recommendations
will be the same as wheat grown for grain.
Forage crop S recommendations
For non-Brassica forage crops it is recommended that on land not receiving regular applications of
organic manure or where sulphur-containing organic materials have not been applied prior to the
growing season, it would be prudent to follow the guidelines proposed by Cussans et al. (2007) for
wheat and apply 20-50 kg SO3/ha in higher risk situations according to the soil type/overwinter rainfall
guidelines.
For forage rape and kale grown on mineral soils, where organic manures have not been applied
regularly in previous years, it is recommended that 50-75 kg/ha SO3 be applied as a sulphate
containing fertiliser in late February to early March.
Forage crop recommendations for other nutrients
Given the lack of available data on the response of forage crops to other nutrients, it is proposed
that no changes be made to the recommendations; apart from introducing fodder beet (roots only)
offtake values, and reducing phosphate and potash recommendations for forage crops grazed in situ
to account for grazing returns.
Options for presenting nitrogen recommendations for grassland
A number of options for presenting grassland N recommendations are presented in Appendix II.
Three options for presenting the N recommendations were considered at the Livestock TWG
Meeting on 26th April 2016 and further guidance will be gathered by AHDB staff at other events
before a final decision is made on the presentation to be used in the “Nutrient Management Guide
(RB209)” in late summer 2016.
67
9.
References
Afzal, M. & Adams, W.A. (1992). Heterogeneity of soil mineral nitrogen in pasture grazed by cattle.
Soil Science Society of America Journal, 56, 1160-1166.
Agricultural and Food Research Council (AFRC) (1993). Energy and protein requirements of
ruminants. An advisory manual prepared by the AFRC Technical Committee on Responses
to Nutrients. CAB International, Wallingford UK.
Aland, A., Lidfors, L. and Ekesbo, I. (2002). Diurnal distribution of dairy cow defecation and
urination. Applied Animal Behaviour Science, 78, 43-54.
Anon. 2014. The British Survey of Fertiliser Practice Fertiliser Use On Farm Crops For Crop Year
2013. https://www.gov.uk/government/collections/fertiliser-usage.
Bailey, J.S. (2016). Managing Sulphur to Optimise Grass Production. Fertiliser Association of
Ireland. Proceedings of Spring Scientific Meeting 2016, 2nd February 2016. Horse and
Jockey, Thurles, Co. Tipperary; pp. 13-16.
Bailey, J.S., Cushnahan A. and Beattie, J.A.M. (1997). Plant and Soil, 197: 137-147
Bailey, J.S., Mellon, J. and K. Hamill, K. (2014). Managing phosphorus to optimise grass
production. Proceedings of Agricultural Research Forum, 2014, p.3 ISBN: 978-1-84170-6054.
Ball, B.C., Bingham, I., Rees, R.M., Watson, C.A. and Litterick, A. (2005). The role of crop rotations
in determining soil structure and crop growth conditions. Canadian Journal of Soil Science,
85, 557-577.
Baumont, R., Michaud, A. & Delaby, L. (2012). Forage services provided by permanent grassland:
grass production and feeding value for ruminants. Fourrages, 219-228.
Bednarek, W., Dresler, S. & Tkaczyk, P. (2015). Nitrogen fractions in timothy grass (phleum
pratense l.) Fertilized with different doses of mineral fertilizers. Journal of Elementology, 20,
49-58.
Betteridge, K., Hoogendoorn, C., Costall, D., Carter, M. and Griffiths, W. (2010). Sensors for
detecting and logging spatial distribution of urine patches of grazing female sheep and
cattle. Computers and Electronics in Agriculture 73, 66-73. Binnie, R.C. and Chestnutt, D.M.B. (1991). Effect of regrowth interval on the productivity of swards
defoliated by cutting and grazing. Grass and Forage Science, 46, 343-350.
Bolland, M.D.A., Allen, D.G. and Barrow, N.J. (2003). Sorption of phosphorus by soils. How it is
measured in Western Australia. Bulletin 4591. Western Australia: Department of Agriculture
and Food.
Borowska, K. & Koper, J. (2011). Dynamics of Changes of Selenium Content in Soil and Red
Clover (Trifolium pratense L.) Affected by Long-Term Organic Fertilization on the
Background of Selected Soil Oxidoreductases. Polish Journal of Environmental Studies, 20,
1403-1410.
68
Boswell, C.C. (1977). Effects of cutting regime on pasture production. New Zealand Journal of
Experimental Agriculture, 5, 403-408.
Brown, L., Scholefield, D., Jewkes, E.C., Preedy, N., Wadge, K., Butler, M. (2000). The effect of
sulphur application on the efficiency of nitrogen use in two contrasting grassland soils. J.
Agric. Sci. 135, 131–138.
Cashman, P.A., McEvoy, M., Gilliland, T.J. and O’Donovan, M. (2016). A comparison between
cutting and animal grazing for dry-matter yield, quality and tiller density of perennial ryegrass
cultivars. Grass and Forage Science, 71, 112-122.
Chadwick, D and Scholefield, D. (2010). Support paper for revision of the RB209 grassland section
- unpublished. North Wyke Research, Devon.
Chaves, B., De Vliegher, A., van Waes, L., Carlier, L. and Marynissen, B. (2009). Change in
agronomic performance of Lolium perenne and Lolium multiflorum varieties in the past 40
years based on data from Belgium VCU trials. Plant Breeding, 128, 680-690.
Ciepiela, G. A., Godlewska, A. & Jankowska, J. (2013). The effect of biostimulant on yields of
mixed grass/red clover stands and chlorophyll content in crop plant leaves under different
nitrogen fertilisation regimes. Fresenius Environmental Bulletin, 22, 3700-3708.
Ciepiela, G.A. and Godlewska, A. (2015). The effect of the biostimulant Kelpak SL on Lolium
perenne L. yield, protein and chlorophyll contents at different nitrogen levels. Agrochimica
59(2), 124-136.
Ciepiela, G.A., Godlewska, A. and Jankowska, J. (2013). The effect of biostimulant on yields of
mixed grassland/red clover stands and chlorophyll content in crop plant leaves under
different nitrogen fertilisation regimes. Fresenius Environmental Bulletin 22(12A), 3700-3708.
Clarke, S., Roques, S., Weightman, R. & Kindred, D. (2016). Modern triticale crops for increased
yields, reduced inputs, increased profitability and reduced greenhouse gas emissions from
UK cereal production. AHDB Cereals & Oilseeds Project Report No. 556.
Collins, J. F. and Cummins, T. (1996). Agroclimatic Atlas of Ireland, Agnet, Dublin, ISBN 0
951155148.
Corcoran, E.M., O’Kiely, P.O., Gilland, T.J., Burke, J.I. and Lynch, M.B. (2016, submitted).
Nitrogen fertiliser rate and timing, and mulch type: effects on forage maize yield and
composition. Submitted to the European Grassland Federation Conference, 2016.
Corrall, J., Morrison, J and Young, J. W. O. (1982). Grass production. In: C. Thomas and J. W. O.
Young (eds.) Milk from Grass, Grassland Research Institute and ICI Agricultural Division, pp.
1-19.
Cougnon, M., Baert, J., Van Waes, C., Poinsard, L. & Reheul, D. (2014). Production of mixed
crops of tall fescue and Italian ryegrass with and without red clover in Belgium. Fourrages,
185-190.
69
Crush, J. R., Ouyang, L. & Nichols, S. N. (2015). Root morphology and architecture, and internal
phosphate use efficiency, in related white clover cultivars of different ages. New Zealand
Journal of Agricultural Research, 58, 302-310.
Curran J., Delaby L., Kennedy E., Murphy J.P., Boland T.M. and O’Donovan M. (2010). Sward
characteristics, grass dry matter intake and milk production performance are affected by pregrazing herbage mass and pasture allowance. Livestock Science 127, 144-154.
Cussans, J.W., Zhao, F.J., McGrath, S.P. & Stobart, R. 2007. Decision support for sulphur
applications to cereals. AHDB Cereals & Oilseeds Project Report No. 419, August 2007,
32pp.
Dale, A. J., Laidlaw, A. S., Bailey, J. S. & Mayne, C. S. (2015). Effect of dairy slurry application rate
and forage type on production, soil nutrient status and nitrogen-use efficiency. Grass and
Forage Science, 70, 44-58.
Dale, A., Aubry, A., Ferris, C., Laidlaw, S., Bailey, J., Higgins, S. and Watson, C. (2013). Critique of
RB209 8th Edition Grassland nitrogen recommendations for dairy systems. AFBI Report.
171pp.
Daly, K.; Styles, D.; Lalor, S.; et al. (2015). Phosphorus sorption, supply potential and availability in
soils with contrasting parent material and soil chemical properties. European journal of soil
science, 66(4), pp.792-801.
Dampney, P.M.R. (1992). The response of whole year grass silage production to nitrogen.
Unpublished report on the analysis of data used to prepare RB209 6th and 7th editions. 28
pp.
Deenen, P. and Middelkoop, N. (1992). Effects of cattle dung and urine on nitrogen uptake and
yield of perennial ryegrass. Netherlands Journal of Agricultural Science, 40, 469-482.
Di, H.J., Cameron, K.C., Silva, R.G., Russell, J.M. and Barnett, J.W. (2002). A lysimeter study of
the fate of N15 labelled nitrogen in cow urine with or without farm dairy effluent in a grazed
dairy pasture soil under flood irrigation. New Zealand Journal of Agricultural Research, 45,
235-244.
Dennis, S.J., Moir, J.L., Cameron, K.C., Di, H.J., Hennessy, D. and Richards, K.G. (2011). Urine
patch distribution under dairy grazing at three stocking rates in Ireland. Irish Journal of
Agricultural and Food Research 50, 149-160.
Durant, D. & Kerneis, E. (2015). Effects of fertilisation on forage yield, feed value, and plant
diversity in a permanent grassland located in the marshes of Western France. Fourrages,
157-165.
Eriksen, J. 2009. Soil sulphur cycling in temperate agricultural systems. Advances in Agronomy,
102, 55-89.
Eriksen, J., Askegaard, M., Rasmussen, J. & Soegaard, K. (2015). Nitrate leaching and residual
effect in dairy crop rotations with grass-clover leys as influenced by sward age, grazing,
cutting and fertilizer regimes. Agriculture Ecosystems & Environment, 212, 75-84.
70
Evans, D.R., Williams, T.A., Jones, S. and Evans, S.A. (1998). The effect of cutting and intensive
grazing managements on sward components of contrasting ryegrass and white clover types
when grown in mixtures. The Journal of Agricultural Science, 130, 317-322.
Frame, J. and Laidlaw, A.S. (2014). "Improved grassland management," The Crowood Press Ltd,
Malborough, UK.
Fried M. and Middelboe V. (1977). Measurement of amount of nitrogen fixed by a legume crop.
Plant Soil 47, 713–715
Ganche, E., O’Donovan, M., Delaby, L., Boland, T. and Kennedy, E. (2012). Effects of grazing
severity on early lactation dairy cow performance. In: Grassland – A European Resource.
Proceedings of the 24th General meeting of the European Grassland Federation, Lublin,
Poland, 3-7 June 2012. Pp 204-206.
Gierus, M., Kleen, J., Loges, R. & Taube, F. (2012). Forage legume species determine the
nutritional quality of binary mixtures with perennial ryegrass in the first production year.
Animal Feed Science and Technology, 172, 150-161.
Godlewska A., Ciepiela G.A. 2016. The effect of the biostimulant Kelpak SL on the content of some
microelements in two grass species. J. Elem., 21(2): 373-381. DOI:
10.5601/jelem.2015.20.2.858.
Grant S.A., Torvell L., Sim E.M. and Small J. (1991). The effect of stolon burial and defoliation
early in the growing season on white clover performance. Grass and Forage Science, 46
173-182.
Haynes, R.J. and Williams, P.H. (1993). Nutrient cycling and soil fertility in the grazed pasture
ecosystem. Advances in Agronomy 49, pp. 119-199.
Higgins, S., Morrison, S. & Watson, C. J. (2012). Effect of annual applications of pelletized
dolomitic lime on soil chemical properties and grass productivity. Soil Use and Management,
28, 62-69.
Hodgson J. and Wilkinson J.M. (1968). The influence of the quantity of herbage offered and its
digestibility on the amount eaten by grazing cattle. Journal of the British Grassland Society
23, 75–80.
Høgh-Jensen H., Schjoerring J.K. (1994). Measurement of biological dinitrogen fixation in
grassland: comparison of the enriched 15N dilution and the 15N natural abundance method
at different nitrogen application rates and defoliation frequencies. Plant Soil 166,153–163.
Hogh-Jensen, H. & Schjoerring, J. K. (2010). Interactions between nitrogen, phosphorus and
potassium determine growth and N-2-fixation in white clover and ryegrass leys. Nutrient
Cycling in Agroecosystems, 87, 327-338.
Humphreys, J., Casey, I. A. & Laidlaw, A. S. (2009). Comparison of milk production from cloverbased and fertilizer-N-based grassland on a clay-loam soil under moist temperate climatic
conditions. Irish Journal of Agricultural and Food Research, 48, 189-207.
71
Humphreys, J., Mihailescu, E. & Casey, I. A. (2012). An economic comparison of systems of dairy
production based on N-fertilized grass and grass-white clover grassland in a moist maritime
environment. Grass and Forage Science, 67, 519-525.
Jafari, A., Connolly, V. and Walsh, E.J. (2003). Genetic analysis of yield and quality in full-sib
families of perennial ryegrass (Lolium perenne L.) under two cutting managements. Irish
Journal of Agricultural and Food Research, 42, 275-292.
Jewkes, E., Blake-Kalff, M. and Blake, L. (2013). Evaluation of the GrowHow N-Min service for use
in maize cropping. Poster presentation at the International Fertiliser Society Annual
Conference, December, 2013.
Kayser, M., Benke, M. & Isselstein, J. (2011). Little fertilizer response but high N loss risk of maize
on a productive organic-sandy soil. Agronomy for Sustainable Development, 31, 709-718.
Keogh, B., McGrath, T. & Grant, J. (2012). The effect of sowing date and nitrogen on the drymatter yield and nitrogen content of forage rape (Brassica napus L.) and stubble turnips
(Brassica rapa L.) in Ireland. Grass and Forage Science, 67, 2-12.
King, C., McEniry, J., Richardson, M. & O'Kiely, P. (2012). Yield and chemical composition of five
common grassland species in response to nitrogen fertiliser application and phenological
growth stage. Acta Agriculturae Scandinavica Section B-Soil and Plant Science, 62, 644-658.
King, C., McEniry, J., Richardson, M. & O'Kiely, P. (2013). Silage fermentation characteristics of
grass species grown under two nitrogen fertilizer inputs and harvested at advancing maturity
in the spring growth. Grassland Science, 59, 30-43.
Knapowski, T., Ralcewicz, M., Spychaj-Fabisiak, E. & Murawska, B. (2012). Effect of the rate of
nitrogen and zinc on the zinc and copper accumulation in grain of spring triticale cultivar
kargo. Journal of Elementology, 17, 421-429.
Korus, A. & Lisiewska, Z. (2009). Effect of Cultivar and Harvest Date of Kale (Brassica Oleracea L.
Var. Acephala) on Content of Nitrogen Compounds. Polish Journal of Environmental Studies,
18, 235-241.
Lynch, J. P., O'Kiely, P. & Doyle, E. M. (2013). Yield, nutritive value and ensilage characteristics of
whole-crop maize, and of the separated cob and stover components - nitrogen, harvest date
and cultivar effects. Journal of Agricultural Science, 151, 347-367.
Malhi S.S., Nyborg M., Harapiak J.T. (1998). Effect of long-term N fertilizer-induced acidification
and liming on micronutrients in soil and in bromegrass hay. Soil and Tillage Research, 48(12): 91-101.
Malhi, S.S., Coulman, B. and Schoenau, J.J. (2009). Maximising timothy forage yield and quality
by balanced nitrogen, phosphorus and sulfur fertilization. Agronomy Journal, 101, 11821189.
Mathot, M., Thelier-Huche, L. & Lambert, R. (2009). Sulphur and nitrogen content as sulphur
deficiency indicator for grasses. European Journal of Agronomy, 30, 172-176.
72
Matthew, C., Hernandez-Garay, A. and Hodgson, J. (1996). Making sense of the link between tiller
density and pasture production. Proceedings of the New Zealand Grassland Association, 57,
83-87.
Mayne, C.S., Patterson, D. and Rankin, W.S.A. (2002). On farm monitoring of grass growth, grass
quality and herbage intake of dairy cows in Northern Ireland. End of Project Technical Report
presented to Agrisearch, July 2002
Michaud, A., Plantureux, S., Pottier, E. & Baumont, R. (2015). Links between functional
composition, biomass production and forage quality in permanent grasslands over a broad
gradient of conditions. Journal of Agricultural Science, 153, 891-906.
Mihailescu, E., Murphy, P. N. C., Ryan, W., Casey, I. A. & Humphreys, J. (2014). Nitrogen balance
and use efficiency on twenty-one intensive grass-based dairy farms in the South of Ireland.
Journal of Agricultural Science, 152, 843-859.
Mihailescu, E., Murphy, P. N. C., Ryan, W., Casey, I. A. & Humphreys, J. (2015a). Phosphorus
balance and use efficiency on 21 intensive grass-based dairy farms in the South of Ireland.
Journal of Agricultural Science, 153, 520-537.
Mihailescu, E., Ryan, W., Murphy, P. N. C., Casey, I. A. & Humphreys, J. (2015b). Economic
impacts of nitrogen and phosphorus use efficiency on nineteen intensive grass-based dairy
farms in the South of Ireland. Agricultural Systems, 132, 121-132.
Minson, D.J. (1990). Forage in Ruminant Nutrition. Academic Press, London.
Morrison, J. (1987). Grassland production: fertilizer-N, water and white clover. In: Nitrogen and
water use by grassland. AFRC Colloquium held at North Wyke Research Station, March
1987. (Ed. R. J. Wilkins). Pp 6-24.Richards, F. J. (1959). A flexible growth function for
empirical use. Journal of Experimental Botany 10, 290-300.
Morrison, J., Jackson, M. V. and Sparrow, P. E. (1980). The response of perennial ryegrass to
fertilizer nitrogen in relation to climate and soil. Report of the joint ADAS/GRI Grassland
Manuring Trial- GM20. GRI Technical Report No 27, 1980.
O’Dovnovan M. and Delaby L. (2008). Sward characteristics, grass dry matter intake and milk
production performance is affected by timing of spring grazing and subsequent stocking rte.
Livestock Science 115, 158-168.
Orr, R.J., Griffith, B.A., Champion, R.A. and Cook, J.E. (2012). Defaecation and urination
behaviour in beef cattle grazing semi-natural grassland. Applied Animal Behaviour Science,
139, 18-25.
Paynter, R.M. and Dampney, P.M.R. (1991). The effect of rate and timing of phosphate fertiliser on
the yield and phosphate offtake of grass grown for silage at moderate to high levels of soil
phosphorus. Grass and Forage Science, 46, 131-137.
PDA (2007a). Results from grass demonstration plots. Potash Development Association leaflet 5b.
PDA (2007b). Nutrient requirements of forage crops. Potash Development Association leaflet 26.
PDA (2008). Forage maize fertiliser requirements. Potash Development Association leaflet 17.
73
Peyraud, J. L., Le Gall, A. & Luescher, A. (2009). Potential food production from forage legumebased-systems in Europe: an overview. Irish Journal of Agricultural and Food Research, 48,
115-135.
Phelan, P., Casey, I. A. & Humphreys, J. (2013a). The effect of target postgrazing height on sward
clover content, herbage yield, and dairy production from grass-white clover pasture. Journal
of Dairy Science, 96, 1598-1611.
Phelan, P., Casey, I. A. & Humphreys, J. (2014). The effects of simulated summer-to-winter
grazing management on herbage production in a grass-clover sward. Grass and Forage
Science, 69, 251-265.
Phelan, P., Keogh, B., Casey, I. A., Necpalova, M. & Humphreys, J. (2013b). The effects of
treading by dairy cows on soil properties and herbage production for three white cloverbased grazing systems on a clay loam soil. Grass and Forage Science, 68, 548-563.
Power, V., Tunney, H. and Jeffrey, D.W. (2005). The phosphorus requirements for silage
production on high fertility soils. Irish Journal of Agricultural and Food Research, 44, 281296.
Remison, S. U., and Snaydon, R. W. (1980). Effects of defoliation and fertilizers on root
competition between Dactylis glomerata and Lolium perenne. Grass and Forage Science, 35,
81-93.
Richards, I.R. (1978). Influence of soil and sward characteristics on the response to nitrogen. In
"Proceedings, International meeting on animal production from temperate grassland, Dublin,
June 1977" (B. Gilsenan, ed.), pp. 45-49, Dublin.
Richards, I.R. and Wolton, K.M. (1976). Spatial distribution of excreta under intensive cattle
grazing. Journal of the British Grassland Society 31, 89-92.
Richards, I.R., Wolton, K.M. and Ivins, J.D. (1976). Note on effect of sheep grazing on yield of
grass swards. Journal of Agricultural Science 87, 337-340.
Ryan, W., Hennessy, D., Murphy, J. J., Boland, T. M. & Shalloo, L. (2011). A model of nitrogen
efficiency in contrasting grass-based dairy systems. Journal of Dairy Science, 94, 10321044.
S.I. No. 610 of 2010. Statutory Instruments. European Communities (Good Agricultural Practice for
protection of waters) Regulations 2010. pp 41-42.
Sampoux, J.P., Baudouin, P., Bayle, B., Béguir, B., Bourdon, P., Chosson, J-F., Deneufbourg, F.,
Galbrun, C., Ghisquière, M., Noël, D., Pietraszek, W., Tharel, B. and Viguié, A. (2011). Field
Crops Research, 123, 117-129.
Schulte, R.P.O., 2007. Modelling phosphorus from grassland: agronomically and environmentally
sustainable advice. End-of-project-report 5188. Teagasc, Johnstown Castle.
Schulte, R.P.O., Melland, A.R., Fenton, O., Herlihy, M., Richards, K. and Jordan, P. (2010).
Modelling Soil Phosphorus Decline: Expectations of Water Framework Directive Policies.
Environmental Science & Policy 13, 472–484. (PDF Download Available). Available from:
74
https://www.researchgate.net/publication/223042325_Modelling_Soil_Phosphorus_Decline_
Expectations_of_Water_Frame_Work_Directive_Policies [accessed Mar 24, 2016].
Scott, N.M., Watson, M.E., Caldwell, K.S. and Inkson, R.H.E. (1983). Response of grassland to the
application of sulphur at two sites in North-east Scotland. J. Sci. Fd. Agric. 34: 357-361.
Scott, P.R. (2010). Nutrient planning in dairy farming on temperate grassland – the need for an
integrated approach. Proceedings of the International Fertiliser Society, 684, 32 pp.
Sinclair, A., Crooks, W. and Coull, M. (2014). Soils information, texture and liming
recommendations. SRUC Technical Note TN656. April 2013, 6pp.
Sinclair, A., Crooks, W., Edwards, T. and Coull, M. (2015a). Management of cobalt in grassland.
SRUC Technical Note TN664. September 2015, 6pp.
Sinclair, A., Edwards, T. and Coull, M. (2015b). Managing soil phosphorus. SRUC Technical Note
TN662. April 2015, 8pp.
Sinclair, A., Shipway, P. and Crooks, W. (2013). Fertiliser recommendations for grassland. SRUC
Technical Note TN652. April 2013, 8pp.
Smith, A., Arnott, R.A. and Peacock, J.M. (1971). A comparison of the growth of a cut sward with
that of grazed swards, using a technique to eliminate fouling and treading. Journal of the
British Grassland Society, 26, 157.
Smith, A., Arnott, R.A. and Macauley, J.R. (1975). Sward productivity under alternate or repeated
cutting of adjacent small areas. Journal of the British Grassland Society, 30, 201-208.
Stevens, R.J. and Watson, C.J. (1986). The response of grass for silage to sulphur application at
20 sites in Northern Ireland. Journal of Agricultural Science (Cambridge), 107: 565-571.
Suter, M., Connolly, J., Finn, J. A., Loges, R., Kirwan, L., SEBASTIA, M.-T. & Luescher, A. (2015).
Nitrogen yield advantage from grass-legume mixtures is robust over a wide range of legume
proportions and environmental conditions. Global Change Biology, 21, 2424-2438.
Suttle, N. F. (2010). Mineral nutrition of livestock. 4th ed. CABI Publishing. ISBN 978-1-84593-4729.
Tallec, T., Diquélou, S., Lemauviel, S., Cliquet, J.B., Lesuffleur, F. & Ourry, A. 2008.
Nitrogen:sulphur ratio alters competition between Trifolium repens and Lolium perenne under
cutting: Production and competitive abilities. European Journal of Agronomy 29, 94–101.
Thomas, C., Reeve, A. and Fisher, G.E.J. (1991). Milk from Grass, 2nd edn. Scottish Agricultural
Colleges, Perth.
Valkama, E., Virkajärvi, P., Uusitalo, R., Ylivainio, K. and Turtola, E. (2016). Meta-analysis of grass
ley response to phosphorus fertilization in Finland. Grass and Forage Science 71(1), 36-53. Webb, J., Fraser, A., Wiltshire, J., Jephcote, C., Rose, R. and Vincent, K. (2015). Review of crop
requirements for sulphur and the adequacy of current sulphur fertilizer recommendations.
Final report for Defra project SCF0308. 51pp.
75
White, S.L., Sheffield, R.E., Washburn, S.P., King, L.D. and Green, J.T. (2001). Spatial and time
distribution of dairy cattle excreta in an intensive pasture system. Journal of Environmental
Quality, 30, 2180-2187.
Wilkins, P.W. (1989). Genotype/harvesting frequency and genotype/nitrogen level interactions for
annual dry matter yield in Lolium perenne in relation to breeding. Euphytica, 41, 207-214.
Wilkins, P.W. and Lovatt, J.A. (2010). Gains in dry matter yield and herbage quality from breeding
perennial ryegrass. In: Grasses for the Future. M. O’Donovan and D. Hennessy (eds.)
Proceedings of an International Conference, Cork, 14-15 October, 2010. Teagasc. 43-50.
Wims, C. M., Delaby, L., Boland, T. M. & O'Donovan, M. (2014). Effect of pre-grazing herbage
mass on dairy cow performance, grass dry matter production and output from perennial
ryegrass (Lolium perenne L.) pastures. Animal, 8, 141-151.
Wojtkowiak, K., Stepien, A., Warechowska, M., Konopka, I. & Klasa, A. (2014). Effect of fertilisation
technique on some Indices of nutritional value of spring triticale grain. Journal of
Elementology, 19, 229-242.
Wortmann, C.S., Dobermann, A.R., Ferguson, R.B., Hergert, G.W., Shapiro, C., Tarkalson, D.D &
Walters, D.T. 2009. High-Yielding Corn Response to Applied Phosphorus, Potassium, and
Sulfur in Nebraska. Agronomy & Horticulture -- Faculty Publications. Paper 327.
Yan, T., Gordon, F.J., Agnew, R.E., Porter, M.G. and Patterson, D.C. (1997). The metabolisable
energy requirement for maintenance and the efficiency of utilisation of metabolisable energy
for lactation by dairy cows offered grass silage-based diets. Livestock Production Science
51, 141-150.
Other reports

AHDB Cereals & Oilseeds project 3699 – ‘Modern triticale crops for increased yields,
reduced inputs, increased profitability and reduced greenhouse gas emissions from UK
cereal production’.

AHDB project 74316 ‘Assessment of varietal characteristics important to low and zero
inorganic nitrogen input herbage production’.

Dale, A., Aubry, A., Ferris, C., Laidlaw, S., Bailey, J., Higgins, S. and Watson, C. (2013).
Critique of RB209 8th Edition Grassland nitrogen recommendations for dairy systems. AFBI
Report. 171pp.

Defra project IF01121 ‘Validation of Fertiliser Manual (RB209) recommendations for
grassland’

Defra project KT018 ‘Nutrient management decision support systems process
improvement’.

Defra project LS3650 ‘Utilise genetic variation within and between improved grass
populations to improve the sustainability of UK grassland’.
76

Wilkins, P.W. and Lovatt, J.A. (2010). Gains in dry matter yield and herbage quality from
breeding perennial ryegrass. In: Grasses for the Future. M. O’Donovan and D. Hennessy
(eds.) Proceedings of an International Conference, Cork, 14-15 October. Teagasc. 43-50.
Acknowledgments

AFBI

BSPB

British Grassland Society

BSPB (British Society of Plants Breeders)

CF Fertilisers UK Ltd

DLF

Ecopt Consultancy

FACTS

Institute of Biological, Environmental and Rural Sciences (IBERS), Aberystwyth University

K+S UK & Eire Ltd

Maize Growers Association (MGA)

NIAB

Potash Development Association (PDA)

Rothamsted Research

Scotland’s Rural College (SRUC)

Sirius minerals (data on sulphur supply to grass)

Yara
77
Appendix I – Search terms and number of results
Titles searched, from 2009 onwards, inverted commas used around search terms. Number of
papers returned refers to pre-screening of titles or abstracts.

Grass and Fertiliser refined by Netherlands or Wales or Germany or England or France or
Denmark or Poland or Ireland = 14 papers

Grass and Fertili* refined by Ireland or Scotland or England or Netherlands or Denmark or
Germany or wales or France = 18 papers

Grass and dry matter yield = 14 papers

Grass and nitrogen refined by Northern Ireland or Wales or Scotland or Ireland or England
= 19 papers

Grass and Potassium = 14 papers

Grass and phosphorus refined by France or Ireland or Scotland or Germany or Netherlands
or Northern Ireland = 14 papers

Grass and sulphur refined by France or England or Belgium = 3 papers

Grass and sulfur refined by France or England or Belgium = 3 papers

Grass and potash = No results

Grass and sward age = 1 paper

Grass and spoilage = No results

Grass and nutrient management = No results

Grass and phosphate = 14 papers refined by Northern Ireland or Scotland or Wales or
England or Ireland

Grass and yield response = No results

Grass and nutrient management = No results

Grass and production refined by Ireland or Wales or Northern Ireland or England = 41
papers

Grass and productivity refined by England or Scotland or Northern Ireland or Ireland or
Wales = 12 papers

Grass and yield response = 9 papers

Grass and livestock refined by Ireland or France or Germany or Belgium = 7 papers

Maize and fertiliser refined by England or Germany or Scotland or Northern Ireland or
France or Denmark or Netherlands or Belgium = 27 papers

Maize and dry matter yield refined by Germany or France or Denmark = 5 papers

Maize and nitrogen refined by Ireland or Wales or Northern Ireland or Scotland or England
= 23 papers

Maize and potassium refined by Scotland or Northern Ireland or Germany = 5 papers
78

Maize and phosphorus refined by Belgium or France or Scotland or Northern Ireland or
Ireland or Germany or Netherlands or England or Denmark = 26 papers

Maize and sulphur = 11 papers

Maize and fertili* refined by Northern Ireland or England or Scotland = 13 papers

Maize and potash = No results

Maize and nutrient management = No results

Maize and phosphate = 8 papers

Maize and nutrient management = No results

Maize and production Ireland or Scotland or England = 15 papers

Maize and yield response = No results

Maize and productivity Germany or Poland or Northern Ireland or England or France or
Denmark or Belgium = 17 papers

Maize and livestock refined by Netherlands = 4 papers

Forage and fertiliser refined by Germany or Switzerland or Denmark = 4 papers

Forage and dry matter yield = 23 papers

Forage and nitrogen refined by Northern Ireland or England or Scotland or Ireland or Wales
= 9 papers

Forage and potassium refined by Switzerland = 1 paper

Forage and phosphorus refined by Netherlands or Switzerland or Germany or Denmark = 4
papers

Forage and sulphur = No results

Forage and sulfur = No results

Forage and fertili* = 9 papers

Forage and potash = No results

Forage and nutrient management = No results

Forage and phosphate = No results

Forage and nutrient management = No results

Forage and production refined by England or France or Northern Ireland or Ireland or
Belgium or Germany or Denmark or Scotland or Wales = 52 papers

Forage and yield response = No results

Forage and livestock refined by Scotland or Germany or Netherlands or England = 8
papers

Triticale and fertiliser refined by Poland = 1 paper

Triticale and fertili* refined by Poland = 6 papers

Triticale and dry matter yields = No results

Triticale and nitrogen = Poland or Denmark = 6 papers

Triticale and potassium = No results
79

Triticale and phosphorus refined by Denmark = 1 paper

Triticale and sulphur = No results

Triticale and sulfur = No results

Triticale and potash = No results

Triticale and nutrient management = No results

Triticale and phosphate = No results

Triticale and nutrient management = No results

Triticale and production = No results

Triticale and productivity = No results

Triticale and yield response = No results

Triticale and livestock = No results

Swede and fertiliser = No results

Swede and fertili* = No results

Swede and dry matter yield = No results

Swede and nitrogen = No results

Swede and potassium = No results

Swede and phosphorus = No results

Swede and sulphur = No results

Swede and sulfur = No results

Swede and potash = No results

Swede and nutrient management = No results

Swede and phosphate = No results

Swede and nutrient management = No results

Swede and production = No results

Swede and productivity = No results

Swede and yield response = No results

Swede and livestock = No results

Kale and fertiliser = 1 paper

Kale and fertiliser = 1 paper

Kale and dry matter yield = No results

Kale and potassium = No results

Kale and phosphorus = No results

Kale and sulphur = No results

Kale and potash = No results

Kale and nutrient management = No results

Kale and phosphate = No results

Kale and nutrient management = No results
80

Kale and productivity = No results

Kale and production = No results

Kale and yield response = No results

Kale and livestock = No results

Clover and fertiliser refined by Scotland or Denmark or Ireland or Northern Ireland Scotland
or Denmark = 3 papers

Clover and fertili* refined by Scotland or Poland or Northern Ireland or Ireland or France or
Denmark or Wales or Belgium = 16 papers

Clover and dry matter refined by Wales = 1 paper

Clover and nitrogen refined by Wales or Denmark or France or Poland or England or
Scotland or Netherlands or Ireland or Germany = 21 papers

Clover and potassium refined by Netherlands or Denmark = 2 papers

Clover and phosphorus = 1 paper

Clover and sulphur refined by France = 3 papers

Clover and potash = No results

Clover and sward age = 1 paper.

Clover and spoilage = No results

Clover and nutrient management = No results

Clover and phosphate = 5 papers

Clover and nutrient management = No results

Clover and production refined by Ireland or North Ireland or England or Belgium or
Denmark or Wales = 14 Papers

Clover and productivity refined by Denmark or France or England = 3 papers

Clover and yield response = No results

Clover and sward age = 1 paper

Clover and livestock = 3 papers

Grassland fertiliser recommendations and farmer attitudes = No results

Nutrient management and grassland farmer = No results

Grassland fertiliser and production = No results

Grassland fertiliser and yield response = 1 paper

Grassland fertiliser and farmer attitudes = No results

Grassland fertiliser and farmer = No results
81
Appendix II – Options for grassland N recommendations
N response Table
A first recommendation table could provide an indication of the dry matter yield produced at
different levels of N fertiliser use for the three Grass Growth Classes at Moderate SNS (Table A1).
The ranges reflect the contrasting amount of grass DM produced by older swards with a low
proportion of modern perennial ryegrass (PRG) cultivars (low productivity) and younger swards
with a high proportion of modern PRG varieties (higher productivity). Indicative DM yield ranges
are based on the three N response curves used in the 8th edition of RB209 (Chadwick and
Scholefield, 2010) and allowing for a DM yield increase of up to 30% due to modern perennial
ryegrass cultivars (Wilkins and Lovatt, 2010). The DM yields assume that swards are cut 4 to 5
times (May-September) and minimal clover content of less than 10-15% cover in mid-season. No
account is taken of in-field losses or spoilage in the clamp. The table provides an indication of the
amount of N fertiliser needed (as crop available N from organic manure or manufactured N
fertiliser) to achieve a given level of grass DM yield. It also indicates that the grass DM yield on
Poor/Very poor GGC sites is likely to be lower than on Good/Very good or Average GGC sites at
any given level of N use; and that at higher N rates the N use efficiency is significantly reduced on
Poor/Very poor GGC sites.
Table A1. Typical ranges for grass dry matter (DM) yield response at different levels of nitrogen (N)
fertiliser use.
Grass DM yield response1 by Grass Growth Class
1
N rate kg/ha
Good/Very Good
Average
Poor/Very poor
0
4-5
3-4
2-3
50
5-7
4-6
3-5
100
6-8
5-7
4-6
150
7-9
6-8
5-7
200
8-10
7-9
6-8
250
9-13
8-12
7-10
300
10-15
9-14
7-11
350
11-16
10-15
8-12
typical whole season DM yield ranges from cut grass swards with minimal clover content over 4-5 cuts. To
account for field losses and spoilage in the clamp, these values can be reduced by 20%.
At the Livestock TWG Meeting on 26th April 2016, Table A1 was thought to be useful in providing
an indication of dry matter yields that can be achieved at different levels of N fertiliser use.
However, it was suggested that the table could be presented in a different colour format from the
82
actual N recommendations, to make it clear that it is a representation of how N response works
and not a recommendation. It was also felt that there was no need to specify that the data is
derived from experiments using 4-5 cuts (see footnote to Table A1).
The N response Table could also be presented as a graph, which more clearly illustrates the
flattening of the N response at higher rates of N, particularly for Poor/Very Poor GGC sites (Figure
A1).
16
Very good/Good GGC
14
Average GGC
Grass DM yield (t/ha)
12
Poor/Very poor GGC
10
8
6
4
2
0
0
50
100
150
200
250
N application rate (kg/ha)
300
350
Figure A1. Grass dry matter (DM) yield nitrogen (N) response curve ranges by grass growth class
(GGC) for grass swards > 3 years old with minimal clover and Low soil nitrogen supply (SNS) to
Moderate SNS.
At the Livestock TWG Meeting on 26th April 2016, Figure A1 was appreciated as a schematic or
illustration of the way the world works in terms of N response. The graph can be used to illustrate
the principles of N response on grassland. Figure A1 could be used to illustrate the principles of
yield response to nitrogen, and the impact of grass growth class (GGC). However, some
considerations needs to be given to the Poor/Very poor GGC especially at high nitrogen rates.
Some land has very limited N response potential due to physical limitations, such as wetness, low
temperature (higher altitude/latitude sites) and north-facing aspect (Figure A1). It was also stated
that it may not be necessary to include both Table A1 and Figure A1 in the recommendations.
83
Grazing N recommendations
An alternative to the current systems approach is to provide recommendations for a narrower
range of circumstances. For example, N recommendations could be provided that assume a low
level of concentrate use and vary the N rate according to stocking rate or indicative DM yield
(Table A2). Such an approach could be used for all livestock types and adjustments can be made
according to the amount of rainfall through the season. The grazing recommendations are
indicative only and provide a base from which to develop a nutrient plan. At the Livestock TWG
Meeting on 26th April 2016, it was stressed that they need to work as a planning tool. It was also
proposed that the N recommendations should be linked to indicative DM yield ranges as the
indictor of intensity, rather than stocking rate (Table A2). The importance of providing supporting
guidance for upland grazing systems and of taking account of the nutrients in applied organic
manures was also highlighted. N should be applied as crop available N from manure applications
(see organic materials section) and/or as manufactured fertiliser, with N rates adjusted through the
season according to summer rainfall and livestock requirements.
Indicative DM yield ranges are based on the N response curve used in the 8th edition of RB209
(Chadwick and Scholefield, 2010) for Very Good/Good Grass Growth Class land and allowing for a
DM yield increase of up to 30% due to modern perennial ryegrass cultivars (Wilkins and Lovatt,
2010). The N rates are selected to avoid high N concentrations in herbage over a 21-30 day
grazing rotation. Grazing rotation length will vary through the season, but in some higher output
systems the average is around 25 days. The recommendations take account of highest growth
rates in late spring and early summer (April to June); the potential for limited rainfall (and soil water
supply) in mid-summer; and high grass growth potential in the autumn (around one third of grass
DM yield can be produced in August to October).
Table A2. Nitrogen (N) recommendations for grazing land by stocking rate.
N application rate (kg/ha)1 per grazing rotation and approximate
application date
Indicative
DM yield
1
Jan/
Feb
Mar
Apr
May
Jun
Jul
Aug
applied
(t/ha)
30
5-7
30
20
6-8
30
30
7-9
40
30
9-12
30
30
302
40
2
40
11-14+
1
(kg N/ha)
4-5
10-13
Total N
30
30
50
20
80
30
30
130
30
30
30
30
180
40
30
30
30
30
230
50
50
40
30
30
270
The recommendations take account of N recycled at grazing
84
2
Only applicable to areas with a long grass growing season; the first N application could be applied as early
as mid to late January, with the second application in early March. No N should be applied after mid-August.
Rates shown apply to:
 Very Good/Good Grass Growth Class and Moderate SNS land. Rates should be adjusted
through the season according to grass growth, summer rainfall and livestock requirements.
 Grazed swards with low clover only (i.e. less than 10-15% cover in mid-season; for grass/clover
swards see ‘Grazing of Grass/Clover Swards – Nitrogen’).
Don’t forget to deduct nutrients applied as organic manures (see organic materials section)
N recommendations by ‘management sequence’ and GGC
Recommendations could be provided that indicate the amount of N to apply at each defoliation and
the amount of grass DM yield that should be expected from low clover swards according to GGC
(Table A3). Indicative DM yield ranges are based on the three N response curves used in the 8th
edition of RB209 (Chadwick and Scholefield, 2010) for Very Good/Good, Average and Poor/Very
poor Grass Growth Class land and allowing for a DM yield increase of up to 30% due to modern
perennial ryegrass cultivars (Wilkins and Lovatt, 2010). The recommendations take account of the
fact that Poor/Very poor GGC sites, on average, need more N to produce a given amount of grass
DM than Average or Good/Very good GGC sites. ‘G’ means a grazing rotation to which N should
be applied, i.e. N is applied prior to the grazing rotation. A single ‘G’ therefore means that N should
only be applied to the first grazing rotation. ‘S’ means a grass silage cut for which a N fertiliser
application should be made, e.g. for a second cut of silage, N is applied immediately after first cut.
A two ‘S’ sequence (S S) means that N fertiliser should be applied in early spring ahead of first cut
and immediately after first cut for a second cut of silage. ‘H’ means a N application associated with
a cut of hay or haylage.
The N recommendations included the following considerations:

Grazing recommendations take account of 15% of N recycled at grazing (rounded up or
down to the nearest 10 kg N/ha).

Grazing - the maximum amount of N per application was set at 80 kg N/ha. This is in line
with previous N recommendations for grazing.

The maximum whole season N rate was set at 330 kg N/ha to reduce risk of exceeding
maximum/optimum N rate and to improve N use efficiency, as indicated by Defra IF01121
N response curves.

To reduce the risk of N fertiliser being applied at rates that exceed the maximum yield
achievable at sites/seasons with limited growth potential:
o
First cut recommended rates were set at a maximum of 120 kg N/ha.
o
Second cut recommended rates were set at a maximum of 90 kg N/ha.
85

o
Third cut recommended rates were set at a maximum of 75 kg N/ha.
o
Fourth cut recommended rates were set at a maximum of 50 kg N/ha.
For grazing and cutting combinations, the N rate for the final grazing was reduced by 15%
to take account of N cycling at the first grazing unless no N was applied to the final grazing
in which case the N rate for the first grazing was reduced.

For hay production, overall N rates were reduced by 5% relative to silage cut yields to take
account of later cutting and bulking out of the hay crop.
Good/Very good GGC sites with 2-10 year old swards are likely to achieve target DM intake values
at the higher end of the range. First year grass leys can achieve yields that are 10-20% above the
upper end of the range. Poor/ very poor GGC sites are likely to achieve DM intake levels towards
the lower end of the range in most years.
At the Livestock TWG Meeting on 26th April 2016, it was stated that the presentation was
understandable and user friendly; and a convenient starting point for discussion of grazing and
cutting intensities. It was agreed that DM yield rather than DM intake should be used (Table A3);
hay recommendations should be integrated into the table; and grass/clover recommendations
should be retained as a separate section. Guidance on optimal timings of N applications in cutting
and grazing situations should also be included.
The recommendations were also trialled at agricultural events such as Scotgrass on 18th May 2016
and Beef expo on 20th May 2016 with a number of respondents stating that total N rates should be
included in the management sequence table; and that the hay recommendations should include an
option for a grazing-hay-grazing sequence (Table A3).
86
Table A3. Grassland nitrogen (N) recommendations by ‘management sequence’, indicative dry matter yields (t DM/ha) and GGC. All rates apply to
Moderate SNS sites with minimal clover content (i.e. less than 10-15% cover in mid-season).
N rates (kg N/ha) by Grass Growth Class
Good/ Very good
Average
Poor/ Very poor
Indicative
DM yield
(t/ha)1
3-4
N rates (kg N/ha)
Total (kg N/ha)
N rates (kg N/ha)
Total (kg N/ha)
N rates (kg N/ha)
Total (kg N/ha)
0
0
30
30
50
50
6-8
30-30-20
80
50-30-30
130
70-70-30
170
9-12+
30-30-30-30-30-30
180
40-40-40-40-30-30
220
N/A3
N/A3
SG
5-7
50-0
50
70-0
70
100-50
150
SG
6-8
60-30
90
70-40
110
N/A2
N/A2
SSG
7-9
70-30-30
130
80-40-40
160
100-75-75
250
Management
sequence
G
GGG
GGGGGG
3
SSG
8-11
80-50-40
170
80-70-50
200
N/A
N/A3
SSSG
10-13
90-60-60-40
250
110-75-75-50
310
N/A3
N/A3
SSSSG
11-14+
100-75-75-30-30
310
100-75-75-50-30
330
N/A3
N/A3
GSG
5-7
40-0-0
40
30-40-0
70
40-70-30
140
GSG
7-9
30-70-30
130
40-80-30
150
80-110-50
240
9-12
40-90-40-30
200
50-100-60-30
240
N/A
N/A3
11-14+
50-100-60-60-30
300
50-100-70-70-30
320
N/A3
N/A3
H
4-5
30
30
50
50
80
80
HG
6-8
60-30
90
80-30
110
100-80
180
GHG
7-9
30-60-30
120
40-80-40
160
50-100-80
230
GSSG
GSSSG
1
For grass swards in their second year or older
Rates required over two or more ‘defoliations’ may impair grass quality
3
DM yield yields unlikely to be achieved
2
Don’t forget to deduct nutrients applied as organic manures (see organic materials section)
87
3
Grass silage recommendations
Nitrogen recommendations should relate to modern grassland systems; provide some indication of
the level of production; and be clear. A significant proportion of farmers and advisers consulted as
part of telephone surveys, email surveys and focus groups in Defra project IF01121 called for a
simplification of the N recommendations. Many farmers and advisers requested explicit guidance
on how much nitrogen to apply prior to each cut of silage to achieve a given level of production.
Table A4 provides N recommendations and associated indicative in-field DM yields for grass silage
production on swards with low clover content over 1 to 4 cuts. Indicative DM yield ranges are
based on the N response curve used in the 8th edition of RB209 (Chadwick and Scholefield, 2010)
for Very Good/Good Grass Growth Class land and allowing for a DM yield increase of up to 30%
due to modern perennial ryegrass cultivars (Wilkins and Lovatt, 2010).
The recommendations are applicable to any livestock system in a Very good/Good GGC and
Moderate SNS situation. Adjustments can be made according to summer rainfall and livestock
requirements. The user has to assess how the recommendations relate to their particular
production system and their overall requirement for grass DM. However, advice on energy
requirements can be provided as part of forage and feed planning recommendations.
Nitrogen rates per silage cut are also provided in Table A3 as part of a ‘management sequence’
approach, but there may be advantages to providing silage crop recommendations for clarity and
ease of use, as is the case for hay N recommendations in the “Fertiliser Manual (RB209)”.
Table A4. N recommendations (kg N/ha) for grass silage by cut.
Indicative DM
yield1 (t/ha)
N application rate (kg N/ha)
1st cut
2nd cut
3rd cut
4th cut
Total N
applied2
(kg N/ha)
5-7
70
-
-
-
70
7-9
80
50
-
-
130
10-13
100
75
753
-
250
11-14+
120
90
703
303
310
1
DM yield as harvested in the field for all cuts combined. Does not include spoilage in the clamp.
2
As manufactured fertiliser and crop available N from organic materials.
3
If previous growth has been severely restricted by drought, reduce or omit this application.

N rates are for Very good/Good GGC sites (i.e. sites that receive adequate rainfall
throughout the summer to support grass growth - see GGC Table) with Moderate SNS. For
High SNS sites, apply 10 kg N/ha less for first cut, and 20 kg N/ha less for second cut. For
Low SNS sites, apply 10 kg N/ha more for first cut, and 20 kg N/ha more for second cut.
88

For 1st cut rates over 90 kg N/ha, apply 40 kg N/ha in mid-February to early March with the
remainder in late March to early April and at least 6 weeks before cutting.

Applications for second and subsequent cuts should be made as soon as possible after the
previous cut.
Don’t forget to deduct nutrients applied as organic manures (see organic materials section)
At the Livestock TWG Meeting on 26th April 2016, it was stated that grass silage production is a
fixed cost activity so it is important to maximise the return on the investment. However, it was also
noted that some livestock farmers (e.g. beef producers) take a later single cut of silage for bulk
feed using moderate N rates and the recommendations should reflect this. The Livestock TWG
thought the recommendations were clear and usable.
Systems approach
The “Fertiliser Manual (RB209)” provided grassland N recommendations for a wide range of
livestock systems. A large number of users were able to match their farm to the recommendation
tables. However, providing a large number of stocking rate – concentrate – milk yield combinations
gave the impression that all situations were covered, when in fact they were not. There was also
an undue level of precision associated with the current recommendations, which were accepting of
production systems that are not profitable under current market conditions (e.g. the use of high
levels of purchased feeds).
It was therefore proposed by the project team that a reduced number of cells could be provided
within the recommendation tables to cover a range of situations (Table A5). A simplified and
revised systems approach could still have provided a useful level of flexibility for multiple
approaches to feeding livestock within dairy, beef and sheep production systems. A flow chart
taking users through the process of generating a recommendation could also have aided clarity
and comprehension. Indicative yield ranges were also included to reflect the difference in
productivity between older and younger swards (Table A5).
The recommendations in Table A5 used the same energy model that underpinned the “Fertiliser
Manual (RB209)” grassland N recommendations due to the uncertainties in changing numerous
factors within the model, including DM yield response to applied N and livestock energy
requirements. The project team proposed that if a detailed systems approach (as used in the 8th
edition) was to be retained, recommendation tables would be provided for dairy, beef and sheep
systems across all three GGC’s.
89
Table A5. Whole season total N requirement for Good / Very good GGC dairy grassland.
Dairy Grass Growth Class Very Good / Good
Total N requirement
Milk yield
Concentrate use
Stocking
Rate
l/cow/yr
t/cow/yr
LU/ha
kg/ha
Indicative yield
(t DM/ha)
kg/ha
3.5
310
10-13
210
3.0
260
9-12
150
2.6
310
10-13
210
2.2
240
9-12
150
2.2
360
11-14
340
1.6
210
8-11
170
1.9
330
10-13
320
1.5
230
9-12
190
2.4
350
11-14
330
1.8
210
9-13
180
2.1
320
11-14
290
1.7
220
9-13
190
2.2
310
11-14
240
2.0
270
10-13
200
8,000 to
10,000
8,000 to
10,000
6,000 to
8,000
3.7
1.5
0.9
4,000 to
6,000
0.9
< 5,000
extended
grazing
Grazed
4.4
6,000 to
8,000
4,000 to
6,000
Cut
0.5
0.5
At the Livestock TWG Meeting on 26th April 2016, it was agreed that the 8th edition systems
approach provided a useful basis for the development of future recommendations, and that it
should be built on to provide recommendations in a different format using the same yield response
model. It was also noted that the principles underpinning the 8th edition recommendations were
well formulated and formed the basis of a useful approach that takes account of a wide variety of
contrasting production systems.
90
‘Road testing’ of grassland N recommendations
Three options for presenting the N recommendations were considered at the Livestock TWG
Meeting on 26th April 2016

A ‘two tables’ option – one for silage and one for grazing (Tables A2 and A4)

A ‘management sequence’ option (Table A3)

Both options (Tables A2, A3 and A4)
A consensus was not reached and it was decided that all three options (along with the N response
schematic) should be ‘road tested’ at agricultural events such as Scotgrass on 18th May 2016 and
Beef expo on 20th May 2016. The overall feedback from farmers and advisers consulted at these
events was positive with suggestions incorporated into Tables A2, A3 and A4. Stakeholders
appreciated the N response schematic and indicated a need for guidance on how to
measure/monitor grass DM yields. The general consensus was that using both options (i.e. all
three tables) was not necessary with approximately 60% favouring the ‘management sequence’
option and 40% favouring the ‘two tables’ option. Further guidance will be gathered by AHDB staff
at other events before a final decision is made in late summer 2016.
91
Appendix III - companies invited to submit data and/or opinions
ADAS
Agrii
AHDB (RB209 Livestock Technical Working Group)
Association of Independent Crop Consultants (AICC; various individual contacts)
Bangor University
British Grassland Society
BSPB (British Society of Plants Breeders)
CF Fertilisers UK Ltd
DLF
Ecopt Consultancy
Frontier Agriculture Ltd
Germinal
Harper Adams University
Hutchinsons
Institute of Biological, Environmental and Rural Sciences (IBERS), Aberystwyth University
International Fertilizer Society
James Hutton Institute
K+S UK & Eire Ltd
Maize Growers Association (MGA)
NIAB
NRM Laboratories
OMEX
Origin Fertilisers
Potash Development Association (PDA)
Rothamsted Research
Scotland’s Rural College (SRUC)
Sirius Minerals
SOYL
Teagasc
Wageningen UR (Netherlands)
Yara
92